Legal Research Memo

Canonical path: skills/legal-research/legal-research-memo/SKILL.md

Agent Trigger Description

Use when producing a structured legal research memo in response to a specific legal question, organizing analysis using IRAC discipline (Question Presented, Brief Answer, Facts, Assumptions, Discussion/Analysis, Conclusion) with explicit sourcing requirements and attorney verification checkpoints.

What this produces: Structured legal research memo using IRAC discipline for attorney review

What you give it: The specific legal question presented; The known facts; The applicable jurisdiction; Any authority already provided

When to use it: A user asks to "research this issue," "write a research memo," "what is the law on X," or "can you analyze whether Y is legal."

At a glance

Practice area	Legal Research
Category	research
Risk level	medium
Recommended quality checks	`attorney-review-gate` `citation-integrity-check` `source-validation-check` `jurisdiction-deadline-gates` `privilege-confidentiality-check` `output-format-compliance-check`
Eval coverage	Manual eval ready
Compatible platforms	chatgpt, claude, cursor, codex, gemini, generic-md
Related skills	brief section drafter, negative treatment check, statutory interpretation, source validation

Example output not yet available.

Selective execution package

This skill has a typed execution contract at skills/legal-research/legal-research-memo/SPEC.json. Its compact core is suitable for routing and quick triage. Standard or deep execution also needs the resources selected by the contract.

Selection rule: choose a declared mode and evaluate each module's machine-readable activation object exactly. Missing activation inputs fail closed. Do not load every conditional module merely because an input is absent.

Declared modes: quick-triage, standard, deep-review

Budget scenarios: 0

Portable copy: “Copy Full Package” includes every possible resource, so it is complete but not context-minimal. Use MCP get_skill_context for the exact selected bundle.

Module	Kind	Activation	Canonical path
`legal-research-memo-template`	template	Required; no conditional activation.	`skills/legal-research/legal-research-memo/templates/legal-research-memo.md`
`citation-type-taxonomy`	reference	Explicit selection only; no machine-readable activation.	`skills/legal-research/references/citation-type-taxonomy.md`
`courtlistener-connector-guide`	connector	Explicit selection only; no machine-readable activation.	`connectors/courtlistener.md`
`negative-treatment-workflow`	workflow-module	Explicit selection only; no machine-readable activation.	`skills/legal-research/negative-treatment-check/SKILL.md`
`source-validation-workflow`	quality-check	Required; no conditional activation.	`skills/legal-methodology/source-validation/SKILL.md`

Typed contract: SPEC.json

{
  "schema_version": "2.0",
  "skill_id": "legal-research/legal-research-memo",
  "gates": {
    "custom": [
      {
        "id": "specific-question-required",
        "condition": "The legal question is vague, compound, or materially underspecified.",
        "action": "stop-and-ask",
        "reason": "The research scope must be confirmed before authorities are gathered or conclusions are drafted."
      },
      {
        "id": "verified-authority-required",
        "condition": "A legal proposition would be stated without user-provided or independently verified authority.",
        "action": "require-attorney-confirmation",
        "reason": "Unverified legal authority must remain a visible placeholder and may not be presented as existing or current law."
      },
      {
        "id": "formal-opinion-boundary",
        "condition": "The requested work product is a formal legal opinion letter or an immediate client-facing legal conclusion.",
        "action": "stop-and-escalate",
        "reason": "Formal opinions and real-time legal advice require direct attorney authorship and professional-responsibility review."
      }
    ]
  },
  "execution_modes": [
    {
      "id": "quick-triage",
      "enabled": true,
      "purpose": "Refine the question presented, identify jurisdiction and factual gaps, propose a research plan, and list authority types to verify without drafting a substantive conclusion.",
      "output_detail": "minimal",
      "quality_checks": [
        "attorney-review-gate",
        "jurisdiction-deadline-gates"
      ]
    },
    {
      "id": "standard",
      "enabled": true,
      "purpose": "Produce the complete attorney-review research memo with verified authority, explicit assumptions, qualified analysis, and open verification items.",
      "output_detail": "standard",
      "quality_checks": [
        "attorney-review-gate",
        "citation-integrity-check",
        "source-validation-check",
        "assumption-audit",
        "legal-prose-polish"
      ]
    },
    {
      "id": "deep-review",
      "enabled": true,
      "purpose": "Run the standard memo plus adverse-authority, currentness, negative-treatment, proposition-support, and citation-integrity passes for each material proposition.",
      "output_detail": "expanded",
      "quality_checks": [
        "attorney-review-gate",
        "citation-integrity-check",
        "source-validation-check",
        "assumption-audit",
        "hallucination-red-team",
        "output-format-compliance-check",
        "privilege-confidentiality-check",
        "legal-prose-polish"
      ]
    }
  ],
  "input_schema": [
    {
      "id": "specific-legal-question-presented",
      "label": "Specific legal question presented",
      "type": "text",
      "required": true,
      "description": "A discrete legal question with confirmed scope.",
      "source_requirement": "user-provided-and-confirmed",
      "may_infer": false,
      "sensitive": false
    },
    {
      "id": "known-facts",
      "label": "Known facts",
      "type": "object",
      "required": true,
      "description": "The legally material facts, each attributable to the user or a provided source.",
      "source_requirement": "user-provided-or-provided-document",
      "may_infer": false,
      "sensitive": true
    },
    {
      "id": "applicable-jurisdiction",
      "label": "Applicable jurisdiction",
      "type": "jurisdiction",
      "required": true,
      "description": "The jurisdiction and body of law governing the question.",
      "source_requirement": "user-provided-and-attorney-confirmed",
      "may_infer": false,
      "sensitive": false
    },
    {
      "id": "authority-already-provided",
      "label": "Authority already provided",
      "type": "document-set",
      "required": false,
      "description": "Cases, statutes, regulations, rules, or secondary sources already supplied for verification and use.",
      "source_requirement": "provided-document",
      "may_infer": false,
      "sensitive": false
    }
  ],
  "output_schema": [
    {
      "id": "structured-legal-research-memo-using-irac-discipline-for-attorney-review",
      "label": "Structured legal research memo using IRAC discipline for attorney review",
      "type": "memo",
      "required": true,
      "description": "Question presented, brief answer, facts, assumptions, issue-by-issue analysis, conclusion, authorities, and open attorney-verification items.",
      "attorney_review_required": true
    },
    {
      "id": "claim-authority-map",
      "label": "Claim-to-authority map",
      "type": "table",
      "required": true,
      "description": "Every material legal proposition mapped to the exact authority and locator that supports it, including verification status and contrary authority.",
      "attorney_review_required": true
    }
  ],
  "evidence_schema": {
    "required_fields": [
      "support_text",
      "authority_status",
      "proposition_supported"
    ],
    "fields": [
      {
        "id": "authority_status",
        "type": "enum",
        "description": "Whether the authority is user-provided, connector-verified, independently verified, secondary-only, or unresolved."
      },
      {
        "id": "proposition_supported",
        "type": "text",
        "description": "The precise proposition for which the source is cited."
      },
      {
        "id": "negative_treatment_status",
        "type": "enum",
        "description": "The recorded treatment status when a negative-treatment check is performed."
      }
    ]
  },
  "modules": [
    {
      "id": "legal-research-memo-template",
      "kind": "template",
      "path": "skills/legal-research/legal-research-memo/templates/legal-research-memo.md",
      "required": true,
      "load_when": "standard or deep-review mode"
    },
    {
      "id": "citation-type-taxonomy",
      "kind": "reference",
      "path": "skills/legal-research/references/citation-type-taxonomy.md",
      "required": false,
      "load_when": "the memo uses multiple authority types or citation forms"
    },
    {
      "id": "courtlistener-connector-guide",
      "kind": "connector",
      "path": "connectors/courtlistener.md",
      "required": false,
      "load_when": "US federal case existence or citation verification is available through CourtListener"
    },
    {
      "id": "negative-treatment-workflow",
      "kind": "workflow-module",
      "path": "skills/legal-research/negative-treatment-check/SKILL.md",
      "required": false,
      "load_when": "deep-review mode or when a material case controls the analysis"
    },
    {
      "id": "source-validation-workflow",
      "kind": "quality-check",
      "path": "skills/legal-methodology/source-validation/SKILL.md",
      "required": true,
      "load_when": "standard or deep-review mode"
    }
  ]
}

Module: legal-research-memo-template

Canonical path: skills/legal-research/legal-research-memo/templates/legal-research-memo.md

Legal Research Memo

Draft legal work product for attorney review. Not legal advice. Do not paste client-sensitive facts into a reusable copy of this template.

PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT


To:	[Supervising Attorney / Matter Team]
From:	[Preparer Name / AgentCounsel workflow]
Date:	[Date of memo]
Re:	[Brief description of legal question and matter name or number]
Matter:	[Matter name / number]
Jurisdiction:	[CONFIRM: jurisdiction and governing law]
Relevant Date:	[Date on which the legal question turns — e.g., date of alleged breach, filing date]

Question(s) Presented

*State each legal question precisely: who did what, under which legal framework, in which jurisdiction, and at what time. Limit each question to one discrete legal issue.*

[Question 1: e.g., "Under [governing law], did [party] [action] so as to [legal consequence] as of [date]?"]
[Question 2, if applicable]
[Add as needed]

Brief Answer(s)

*Direct answer — yes, no, probably, or uncertain — with a one-to-two sentence reason. Reserve analysis for the Discussion section. If the answer turns entirely on unverified facts or law, say so.*

[Yes / No / Probably / Uncertain]. [One-to-two sentence reason. E.g., "Probably yes, because [key rule] appears to apply to [key fact], though [key variable] remains unconfirmed and is a material condition of this answer."]
[Answer to Question 2, if applicable]
[Add as needed]

Facts

*Set out only the legally relevant facts as provided by the user or client. Do not add, infer, or embellish. If a fact is an inference rather than a stated fact, label it as such. Note any factual gaps that may be legally material.*

[Fact 1 — as provided]
[Fact 2 — as provided]
[Inference: [describe inference] — noted as inference, not established fact]
[CONFIRM: [describe factual gap] — legally material; obtain before relying on this memo]
[Add as needed]

Assumptions

*List every assumption made in the analysis — about facts, the applicable legal standard, procedural posture, or authority. Each assumption is a potential point of failure and an attorney verification item.*

[Assumption 1: e.g., "The agreement referenced is governed by [state] law as stated in the governing-law clause; this has not been independently verified."]
[Assumption 2: e.g., "The relevant date for purposes of the statute of limitations analysis is [date]; attorney should confirm which event triggers the limitations period."]
[Assumption 3: e.g., "No amendments to the referenced statute have been enacted between [date] and the date of this memo; attorney should confirm current version."]
[Add as needed]

Discussion / Analysis

*Use IRAC structure (Issue, Rule, Application, Conclusion) for each legal question or sub-issue. Cite the source of every legal rule stated. If no verified source is available, use the [CONFIRM] placeholder.*

Issue 1: [Restate the first legal sub-question]

Rule

[State the applicable legal rule or standard. Identify the source precisely: case name, court, year, and specific holding; or statute/regulation with section number and version. Example: "Under [Statute], § [X], a party must [requirement]. [Citation / source]. If the authority cannot be verified in this session, write: [CONFIRM: rule — no verified authority available for this proposition; attorney must supply and verify before relying on this analysis.]"]

Application

[Apply the rule to the facts stated above. Be explicit about which facts do — and do not — satisfy each element of the rule. Flag disputed or assumed facts. Note where the analysis depends on a factual question that cannot be resolved from the record provided. Do not advocate silently; if the analysis favors one side, say so.]

Element 1 ([element name]): [Does the fact pattern satisfy this element? Why or why not?]
Element 2 ([element name]): [Analysis]
Element 3 ([element name]): [Analysis — if uncertain: [CONFIRM: whether [fact] satisfies this element]]

Conclusion on Issue 1

[Tentative conclusion with explicit confidence level, e.g., "Likely satisfied," "Uncertain — turns on the disputed [fact]," "Unsatisfied on current facts."] [One sentence identifying the key variable on which this conclusion could change.]

Issue 2: [Restate the second legal sub-question, if applicable]

Rule

[Rule with citation / source, or [CONFIRM] placeholder]

Application

[Application to facts, with element-by-element analysis where applicable]

Conclusion on Issue 2

[Tentative conclusion with confidence level]

*[Add additional Issue sections as needed.]*

Adverse Authority and Counterarguments

*Identify legal authority or arguments that cut against the analysis above. Do not omit adverse authority. If no adverse authority has been identified, state that and note it as an attorney verification item.*

[Adverse authority or counterargument 1, with citation / source, and brief note on why it does or does not control]
[CONFIRM: attorney should independently search for adverse authority not identified in this session before relying on this memo]

Conclusion

*Summarize the overall answer to each Question Presented in two to five sentences. Identify the key legal and factual variables on which the answer depends. Do not overstate certainty. Note what would change the answer.*

[Question 1]: [Summary answer with confidence level and key dependencies.]

[Question 2, if applicable]: [Summary answer with confidence level and key dependencies.]

Key variables: [List the two to four facts or legal determinations that most affect the outcome. Attorney should focus verification effort here first.]

Authorities Cited

*List every legal authority referenced in this memo. Every authority must have a source. Authorities that cannot be verified in this session are flagged in the Verified column — attorney must confirm before relying on this memo.*

#	Authority	Type	Citation	Proposition asserted	Source	Verified by attorney
1	[Case / Statute / Regulation / Secondary Source name]	[Case / Statute / Reg / Secondary]	[Full citation]	[Specific holding or rule this authority supports]	[User-provided / Researched — identify source]	- [ ]
2						- [ ]
3						- [ ]
4						- [ ]
5						- [ ]

Authority source key:

User-provided — document or citation supplied directly by the user or client.
Researched — retrieved from a named research source during this session; identify the source.
Connector-verified — citation resolved through a documented connector (e.g., CourtListener — see connectors/courtlistener.md); record the verified opinion URL in the Source column. The citation form is confirmed; whether the cited authority supports the asserted proposition remains an attorney verification item.
[VERIFY-CITE] / [CONFIRM] — authority has not been verified; attorney must locate, review, and confirm before relying on this memo.

Open Items / Attorney Verification

*Checklist of every factual gap, unverified authority, jurisdictional question, ambiguity, or strategic judgment that requires attorney review before this memo is relied upon. Resolve all items before circulation.*

[ ] Confirm jurisdiction and governing law are correctly identified for this matter.
[ ] Confirm the relevant date (triggering event, limitations period start, statute effective date) is correct.
[ ] Verify all case citations exist and that the holdings attributed to them are accurate.
[ ] Verify all statutes and regulations are in their current version and say what this memo claims.
[ ] Independently search for adverse authority not identified in this session.
[ ] Resolve all [CONFIRM: ...] placeholders throughout the memo.
[ ] Confirm all facts stated in the Facts section are accurate and complete.
[ ] Resolve factual gap: [describe gap from Facts section, or delete if none]
[ ] Confirm assumption: [describe assumption from Assumptions section, or delete if none]
[ ] Assess conflict-of-laws or choice-of-law issues, if implicated.
[ ] Assess whether any procedural deadline is affected by this legal analysis.
[ ] Confirm that the privilege and confidentiality designations are appropriate for how this memo will be circulated.
[ ] Attorney to confirm overall conclusions and confidence levels before client communication or reliance.

*This memo is draft legal work product prepared for attorney review. It is privileged and confidential. It is not legal advice and does not establish an attorney-client relationship. No conclusion in this memo should be communicated to or relied upon by any client without attorney review and approval.*

Module: citation-type-taxonomy

Canonical path: skills/legal-research/references/citation-type-taxonomy.md

Shared reference material supporting the AgentCounsel legal-research and legal-methodology skills, used to help produce draft legal work product for attorney review — not legal advice.

Citation Type Taxonomy

This reference gives a reviewer a vocabulary for classifying what *kind* of citation a passage in a draft actually is, before deciding what verifying it would require. Different citation forms carry different verification needs and different hallucination risks — an id. is only meaningful in relation to the citation before it, while a full case citation can be checked in isolation. This taxonomy does not verify anything itself; it tells a skill which question to ask for each citation form it finds.

Attribution. The set of citation-type categories below is adapted from the classification scheme used by eyecite, an open-source citation-extraction library maintained by the Free Law Project, licensed BSD-2-Clause. The descriptions of each type, the verification needs, and the hallucination-risk notes are written independently for AgentCounsel; they are not eyecite's code, docstrings, or documentation text.

How Skills Use This File

skills/legal-methodology/citation-integrity-check/SKILL.md and skills/legal-methodology/source-validation/SKILL.md classify each citation in a draft before assessing its source status or integrity. This taxonomy is the vocabulary for that classification step — it identifies *what form* a citation takes; the two skills above still own the questions of *whether it is real*, *whether it is complete*, and *whether it supports the proposition asserted*. Legal-research skills that assemble authority-heavy drafts (for example skills/legal-research/legal-research-memo/SKILL.md and skills/legal-research/authority-synthesis/SKILL.md) can also use this taxonomy while drafting, to keep short-form citations correctly anchored to the full citation they depend on.

Citation Types

Full case citation

What it looks like: A complete citation to a court decision — case name, reporter volume, reporter abbreviation, starting page (and pin cite if applicable), court, and year, e.g. a pattern like Party A v. Party B, 123 F.3d 456, 460 (9th Cir. 1999).

What a verifier needs: The citation is self-contained — no earlier citation in the document is required to resolve it. Verification means confirming the reporter abbreviation exists (see connectors/reporters-and-courts.md), that the court abbreviation matches a real court, and that the case can be located under that name and citation.

Hallucination risk: This is the highest-risk form, because a fabricated full citation can look completely well-formed — a plausible party-name pattern, a real reporter abbreviation, and a page number within a plausible range, none of which make the case real. A citation "looking right" is not evidence it exists.

Short-form case citation

What it looks like: An abbreviated citation used after a case has already been cited in full, typically the case name plus a shortened reference to the reporter and page, e.g. Party A, 123 F.3d at 462.

What a verifier needs: The preceding full citation for the same case somewhere earlier in the document. A short form cannot be verified in isolation — it is only correct if it actually matches an antecedent full citation for the same case in the same document.

Hallucination risk: A short form can silently drift from its antecedent — citing a different page, or a short form with no matching full citation anywhere in the document at all (an orphaned short form), which is itself a sign the full citation was never actually verified or was dropped in editing.

Supra citation

What it looks like: A citation using "supra" to refer back to a source cited earlier by name, without repeating the reporter citation, e.g. Party A, supra, at 462 or Smith, Legal Treatise, supra note 4, at 12.

What a verifier needs: The specific earlier citation or footnote the supra reference points to. A supra reference is meaningless without resolving exactly which prior citation it targets — this often requires matching a footnote number or a uniquely identifying short form.

Hallucination risk: A supra reference to a footnote or citation number that does not exist in the document, or that points to the wrong source once resolved, silently misattributes a proposition to a citation that never actually supports it.

Id. citation

What it looks like: A citation of Id. or Id. at [page], referring to the same source as the immediately preceding citation.

What a verifier needs: The immediately preceding citation in the document — not the most recent citation of that source anywhere in the draft, but the one immediately before it. An Id. is only correct if nothing else was cited in between.

Hallucination risk: Id. is the form most vulnerable to silent drift during editing — inserting, deleting, or reordering a citation elsewhere in the document can leave a surviving Id. pointing to the wrong source without any visible change to the Id. text itself. Every Id. must be re-checked against its immediately preceding citation whenever the surrounding text is edited.

Statutory citation

What it looks like: A citation to a statute or code section, e.g. a pattern like 15 U.S.C. § 78j(b) or a state-code citation, typically title or chapter, section symbol, and section number.

What a verifier needs: Confirmation of the code, title/chapter, and section number, and — critically — the version or effective date, since statutory text changes over time and a citation with no version marker is ambiguous about which version of the law is being cited.

Hallucination risk: A plausible-looking but non-existent section number, or a real section number whose current text has been amended since the date implicitly assumed by the draft. A statutory citation without an "as of" date is not verifiable as currently stated law without independent research.

Regulation citation

What it looks like: A citation to a codified administrative regulation, e.g. a pattern like 17 C.F.R. § 240.10b-5, or a citation to an agency's Federal Register document by volume and page.

What a verifier needs: The same version-date discipline as a statutory citation, plus confirmation of which agency issued it — regulations are amended, and a codified regulation citation with no effective-date anchor cannot be assumed current.

Hallucination risk: A fabricated part or section number under a real title, or a citation to a rule that has since been superseded, withdrawn, or stayed without the draft acknowledging that possibility.

Law-journal citation

What it looks like: A citation to a law review or journal article, e.g. a pattern like Author Name, Article Title, 100 Some L. Rev. 1 (2020).

What a verifier needs: The author, title, journal volume/abbreviation, starting page, and year — all self-contained, similar to a full case citation, but for secondary rather than primary authority. Confirming existence does not confirm the article actually supports the proposition cited to it.

Hallucination risk: A fabricated author, title, or journal issue that reads as entirely plausible — law review citations are especially easy to invent convincingly because there is no small, checkable universe of reporters or courts to match against, unlike case citations.

Unknown / partial citation

What it looks like: A citation fragment that does not resolve cleanly into any of the categories above — a case name with no reporter cite, a section reference with no code identified, a citation copied with a typo, or a citation abbreviated so heavily that its type cannot be determined from the text alone.

What a verifier needs: More context before any verification step is even possible. Do not guess at the missing elements or "complete" the citation from model background knowledge — treat it as an incomplete citation and flag it for the source to be supplied or the citation corrected.

Hallucination risk: The temptation to silently "fill in" a plausible missing reporter, section number, or year is the single highest-risk moment in citation handling. An unknown/partial citation must never be completed from memory; completing it is how a hallucinated citation gets introduced.

Verification Workflow Per Type

Citation type	Resolves independently?	What must be checked before verification is even possible	Primary hallucination risk
Full case citation	Yes	Reporter and court abbreviation exist (`connectors/reporters-and-courts.md`); case locatable under that name and cite	Entirely fabricated but well-formed citation
Short-form case citation	No — needs its antecedent full citation	Locate and match the earlier full citation for the same case	Drift from, or no match to, the antecedent
Supra citation	No — needs the specific earlier reference	Locate and match the exact footnote/citation the supra points to	Wrong or missing target reference
Id. citation	No — needs the immediately preceding citation	Confirm nothing else was cited in between	Silent drift after document edits
Statutory citation	Yes, with a version caveat	Code, section number, and version/effective date	Fabricated section number; stale version
Regulation citation	Yes, with a version caveat	Title, part/section, issuing agency, effective date	Superseded, withdrawn, or stayed rule cited as current
Law-journal citation	Yes	Author, title, journal/volume, page, year	Fabricated author, title, or issue
Unknown / partial	No	Additional source material or correction from the user	Silently "completing" the citation from memory

Reviewer Notes

This taxonomy classifies citation *form*. It does not substitute for the source-status classification in skills/legal-methodology/source-validation/SKILL.md or the integrity checks in skills/legal-methodology/citation-integrity-check/SKILL.md — use this file to identify the type first, then apply those skills' workflows.
For existence checks on case-law reporter abbreviations and court names, see connectors/reporters-and-courts.md. Existence of a reporter or court abbreviation is not existence of the case itself — see that connector's scope and limits.
An unresolved short-form, supra, or Id. citation (no matching antecedent found in the document) is itself a defect to flag — do not assume the antecedent existed and was simply omitted from view.

Module: courtlistener-connector-guide

Canonical path: connectors/courtlistener.md

CourtListener

Reference material supporting the AgentCounsel skill library, used to help produce draft legal work product for attorney review — not legal advice.

This connector points at CourtListener, the Free Law Project's open database of US case law. It is the first concrete connector under connectors/; the broader framing is in README.md.

The Free Law Project explicitly frames the Citation Lookup endpoint documented here as a guardrail "to help prevent hallucinated citations" (source: Citation Lookup and Verification API). That is the use case this connector enables for an AgentCounsel skill — closing the gap between "I left a [VERIFY-CITE: ...] placeholder" and "I confirmed the case exists with that citation in a reliable corpus."

1. Source

Publisher: Free Law Project (a US-based nonprofit). CourtListener is a research-and-disclosure project, not a commercial legal research vendor.
Cost: Free for read access. No API key required to begin; an account is recommended for monitoring and higher tiers.
API version: Current is v4 (v4.4 at the time this doc was written). The base URL is https://www.courtlistener.com/api/rest/v4/. Older versions exist but should not be relied on for new work; consult the V4 Migration Guide before using anything older. [CONFIRM: API version currency before integrating in a long-lived tool].
Rate limits (per the Free Law Project's published limits at the time this doc was written): up to 5,000 requests per hour for unauthenticated users; 5,000 requests per day for free accounts; the API documentation also notes a 60-requests-per-minute throttle. Limits are noted as in flux pending a new membership model. [CONFIRM: current rate limits before relying on a sustained workflow]. Source: REST API overview and the Free Law Project's discussion on limits.
License of returned content: CourtListener opinions are US public-domain court records; the Free Law Project's own metadata and database are licensed by them under permissive terms. [CONFIRM: applicable Free Law Project terms before redistribution].

This connector documents the free, mostly no-key surface. CourtListener also exposes PACER-derived data and paid features; both are out of scope here.

2. In scope — what CourtListener can verify

CourtListener's corpus and the surface this connector documents support verifying:

US Supreme Court opinions — full historical corpus.
US Courts of Appeals — federal appellate decisions across all circuits.
US District Courts — selective coverage of district-court opinions.
State courts of last resort — varying coverage by state; many state supreme courts are present, with coverage lag and gaps that vary by jurisdiction. Treat state coverage as partial unless a specific state has been confirmed.
Bankruptcy, tax, and specialty federal courts — partial coverage.
Per-opinion basics — case name, citation(s), court, date, judges, and full opinion text where the opinion is published.
Citator-style relationships — what other opinions cite a given opinion (the "OpinionsCited" graph).
Reverse citation lookup — given a citation string, locate the matching opinion(s). The Citation Lookup endpoint parses citations using Eyecite, which the project says was developed against more than 50 million citations going back more than two centuries.

What "verify" means here: confirm that the cited case exists, that the citation form matches what CourtListener has on record, and that the public URL for the opinion can be recorded for attorney review. Verifying that the opinion stands for the proposition for which it is cited is not the connector's job — that is attorney work.

3. Out of scope — what CourtListener does not cover

A skill must keep its existing placeholder discipline for anything in this list:

Many state intermediate appellate and trial court opinions — coverage is partial and varies by state.
Unpublished or sealed proceedings — not in the public corpus.
Very recent decisions — coverage lag varies; new opinions appear on a delay.
Secondary sources — treatises, restatements, law review articles, practice guides.
Statutes, regulations, court rules, and agency guidance — CourtListener's primary focus is case law. Statutory verification belongs to a different connector.
Foreign and international tribunals — out of scope. Use a jurisdiction-specific connector when one is added.
Pinpoint accuracy of internal page numbers in older volumes — verify pinpoint cites against the official reporter.

If a skill needs to verify any of the above, treat the connector as unavailable for that placeholder and retain the existing [VERIFY-CITE: ...] / [CONFIRM: ...] flag.

4. Surface — where to hit

4a. Public web search (no API)

The simplest surface, available in any browser or HTTP-capable tool:

Search URL pattern: https://www.courtlistener.com/?q=<URL-encoded-query>
Opinion page URL pattern: https://www.courtlistener.com/opinion/<opinion-id>/<slug>/

Use this surface for: quick visual confirmation of a case, sharing a stable URL with the attorney, or environments without HTTP tooling.

4b. Citation Lookup API — the primary verification surface

The most reliable way to resolve a citation string to a specific opinion:

Endpoint: https://www.courtlistener.com/api/rest/v4/citation-lookup/
HTTP method: POST (not GET).
Input modes (per the Citation Lookup docs): the API can look up either a single individual citation or parse and look up every citation in a block of text.
Hard limits: at most 250 citations per request (excess parsed but returned with a status value indicating "Too many citations requested"); at most 64,000 characters of text per request.
Response shape (per the official docs): each match returns a citation (the looked-up form), normalized_citations (canonical form(s); multiple entries when ambiguous), start_index and end_index (positions of the citation in the input text), and a status field (200 indicates a found and valid match).
Authentication: uses the same auth and serialization as the rest of the CourtListener API; works unauthenticated subject to the rate limits in Section 1.

This is the endpoint a skill should hit when resolving a [VERIFY-CITE: ...] placeholder. The Eyecite-backed parsing handles ambiguous and non-canonical citation forms that a plain string match would miss.

4c. Search API

For richer queries (court filter, date range, case-name search):

Endpoint: https://www.courtlistener.com/api/rest/v4/search/
HTTP method: GET with query parameters, e.g., ?q=<query>.
Supported filters (consult the live Search API docs for the current list): query text, document type, court filter, date filed range. [CONFIRM: exact filter parameter names against the live docs — they change with API minor versions]. The Free Law Project recommends sending an OPTIONS request to any endpoint to discover the current filter set.

Use this for: locating a case by partial information when only a name or topic is known, narrowing within a specific court or date window.

4d. Opinion fetch

For pulling the full text of a verified opinion:

Endpoint pattern: https://www.courtlistener.com/api/rest/v4/opinions/<opinion-id>/
HTTP method: GET.
Use for: confirming that a paraphrase or quotation matches the source text. Confirming the quotation exists in the opinion is a structural check; the legal weight of the quoted passage remains an attorney determination.

4e. MCP tool surface

When the user's environment includes an MCP server that wraps CourtListener (community wrappers exist; check the user's installed MCP servers), prefer the MCP tool surface over raw URL fetches — it handles authentication, retries, and rate-limit backoff. The connector contract is the same: query, record the verified URL, fall back to placeholder if unavailable. This connector does not configure or install any MCP server — see connectors/README.md.

5. Calling pattern from a skill

When a skill's Workflow reaches a citation it would otherwise mark [VERIFY-CITE: ...] or [CONFIRM: ...], and the citation is in scope under Section 2:

Look up the citation. For a citation string in a draft, POST to the Citation Lookup endpoint (4b). For a case-name or topic-based lookup, use the Search endpoint (4c).
Branch on the result:
- Exact match (single result): Record the case in the Authorities Cited table (or equivalent) with the CourtListener opinion URL appended, e.g., Erie R.R. Co. v. Tompkins, 304 U.S. 64 (1938) — verified via CourtListener: https://www.courtlistener.com/opinion/<id>/erie-railroad-co-v-tompkins/. Replace the [VERIFY-CITE: ...] placeholder with [ATTORNEY TO CONFIRM: proposition supported by the cited case] — the citation now exists; whether it stands for the asserted proposition is still attorney work.
- Multiple matches: Do not pick one. Record all matches, retain the [VERIFY-CITE: multiple matches in CourtListener — attorney to select] flag, and escalate.
- No match: Retain a [VERIFY-CITE: not found in CourtListener — attorney to verify in another source] flag. Do not delete the asserted citation; the attorney needs to see what was claimed.
- Out of scope (per Section 3): Keep the original placeholder. Do not query CourtListener for state intermediate appellate cases or non-case-law authority and treat absence-of-result as a verification.
If the connector is unavailable (no MCP tool, no HTTP capability, rate-limit exceeded): retain the original placeholder unchanged. Add [VERIFY-CITE: not verified — no CourtListener connector available in this session] so the attorney sees the gap.

The Output Format of any skill that consumes this connector should record the verification source per authority — see the per-skill changes in skills/legal-research/legal-research-memo/SKILL.md, skills/litigation/brief-section-drafter/SKILL.md, and skills/litigation/claim-chart/SKILL.md.

6. Fallback behavior

Outcome	Action
Single match	Record verified URL; mark `[ATTORNEY TO CONFIRM: proposition supported]`
Multiple matches	List all; mark `[VERIFY-CITE: multiple matches — attorney to select]`
No match	Mark `[VERIFY-CITE: not found in CourtListener — attorney to verify elsewhere]`
Out of scope per Section 3	Retain the original placeholder; do not query
Connector unavailable	Retain the original placeholder; add `not verified — no CourtListener connector` note
Rate-limit reached	Treat as unavailable for the remainder of the session; flag in Open Items

7. Limits and known failure modes

Coverage lag. New opinions appear with a delay. A citation that does not appear today may appear next week. Do not infer that absence in CourtListener means a case was not decided.
State coverage is partial. A negative result for a state court should be read as "not in CourtListener," not as "does not exist."
Citation form drift. Older citations, parallel citations, and unconventional formats can return false negatives. The Citation Lookup endpoint's Eyecite parser handles many such forms; for stubborn cases, try both the volume-page form (304 U.S. 64) and the case name (Erie v. Tompkins).
Pinpoint citations. Confirming the case exists is not confirming a pinpoint page or paragraph. Pinpoints remain attorney-verification items.
Cited-but-not-quoted propositions. A case being real does not mean the proposition asserted in the draft is supported by the case. That determination remains with the attorney.
Versioning. Opinions can be revised; en banc rehearings and amended opinions can change pin cites. Record the date of the verification and re-verify before reliance on a sensitive citation.
API surface drift. The Free Law Project is actively developing v4; minor versions add and rename filters. Treat the URLs and parameter names above as the shape *at the time this doc was written* and re-check the live docs (/help/api/rest/) before automating against them.

8. What this connector does not do

Does not verify statutes, regulations, court rules, or secondary sources.
Does not validate that a case is still good law (Shepardizing / KeyCiting is a separate operation that this free surface does not perform).
Does not characterize a holding, the level of generality of a holding, or the precedential weight of a case in a particular court.
Does not assess whether a case is binding, persuasive, or distinguishable in the operative forum.

All of those remain attorney-verification items. The connector closes the "does this case exist with this citation" question. Nothing more.

Module: negative-treatment-workflow

Canonical path: skills/legal-research/negative-treatment-check/SKILL.md

name: Negative Treatment Check description: "Use when you need to check whether the authorities cited in a draft or research memo are still good law — organizing a citator-style verification plan for each authority, recording provided citator-report signals as attributed claims, and classifying each authority's verification status, without ever asserting from model memory that a case is good law, overruled, or distinguished." practice_area: legal-research task_type: analysis jurisdictions: [] risk_level: high requires_attorney_review: true inputs:

"The draft, memo, or citation list whose authorities need treatment verification"
"For each authority: the full citation as it appears in the source"
"Optional: any citator reports or validation printouts the user has already obtained"
"Optional: access to a citation-verification connector (e.g., connectors/courtlistener.md) for federal case law"

outputs:

"Authority-by-authority treatment-verification table"
"A verification plan for each authority, using available connectors and user-supplied citator reports"
"Verification-status classification and attorney sign-off items"

related_skills:

skills/legal-research/authority-synthesis/SKILL.md
skills/legal-research/legal-research-memo/SKILL.md
skills/legal-methodology/citation-integrity-check/SKILL.md

tags:

legal-research
citator
negative-treatment
good-law
citation-verification
source-validation

Negative Treatment Check

Purpose

Organize a disciplined "is this still good law" check for every authority cited in a draft or research memo. For each authority, the skill records the citation as provided, structures a verification plan using available connectors and any user-supplied citator reports, and classifies the authority's verification status. It never asserts from model memory that a case is good law, has been overruled, superseded, limited, or distinguished — model-memory treatment claims are exactly the hallucination this library exists to prevent. Every treatment conclusion is either drawn from a provided, attributed source or left as an explicit [Verify current law] item for attorney confirmation. This is draft legal work product for attorney review, not legal advice.

Use When

A research memo or brief is nearly final and every cited authority needs a good-law check before it is filed or relied upon.
A user asks "is this case still good law?" or "have any of these authorities been overruled?" and wants the verification organized rather than guessed.
A citation list has been assembled and the team needs a treatment-verification plan before an attorney signs off.
Citator reports have been pulled from a vendor service and the signals need to be organized into a per-authority status table.
A cited authority is old, from a fast-moving area, or central to the argument, and its current status must be confirmed.

Required Inputs

The draft, memo, or citation list whose authorities need checking, with each authority's full citation as it appears in the source. If only a description is provided, request the actual citations — never reconstruct a citation.
Optional: any citator reports or validation printouts the user has already obtained from a citation service. These are recorded as attributed claims from a provided source.
Optional: access to a citation-verification connector (for example, connectors/courtlistener.md for federal case law) — used to confirm that an authority exists and to retrieve its text and later history where the connector supports it.

If no citations are provided, stop and request them. Do not generate or assume citations to check.

Do Not Use When

The task is to synthesize authorities into a rule or to build the substantive analysis — use authority-synthesis; run this check after synthesis, before finalizing.
The task is to verify that citations are formatted correctly and that quotations match their sources — use citation-integrity-check; this skill checks treatment (still-good-law status), not citation form.
The task is to draft the research memo itself — use legal-research-memo.
The user wants a definitive statement that a case is or is not good law without any provided citator source — this skill cannot supply that; it organizes verification and flags the gap.
No authorities are cited — there is nothing to check.

Legal Safety Rules

Source and citation discipline. Follow core/source-and-citation-discipline.md. Never invent legal authority, citations, quotations, statutes, cases, regulations, filing deadlines, or procedural rules. Label what is a provided source, a user-provided fact, an assumption, a legal inference, or an item requiring attorney verification, and use a citation placeholder such as [Attorney to insert authority] when no source is available.
Produce draft legal work product for attorney review. This is not legal advice.
Never assert treatment from model memory. Do not state — from your own knowledge — that any authority is good law, has been overruled, reversed, superseded, abrogated, limited, criticized, or distinguished. Treatment claims not drawn from a provided, attributed source are the exact hallucination risk this skill exists to contain. State this limitation plainly in the output.
Record vendor citator signals as attributed claims, never as confirmed facts. When a user supplies a citator report, record what the report states (its flags, its listed citing references, its date) attributed to that report — never independently "confirm" or upgrade it. Note that citator services can lag and can miss treatment; the attorney verifies.
Connector results verify existence and history, not legal status. A connector can confirm an opinion exists, retrieve its text, and surface later citing documents where supported — it does not adjudicate whether the authority remains good law. Frame connector findings accordingly.
Classify every authority's status explicitly, and default to [Verify current law] whenever the status rests on anything other than a provided, attributed citator source.
Never compute or assert a date. Record the dates a provided report or connector states, attributed to the source.
Distinguish throughout: what a provided report or connector states (attributed), what the user stated, what is assumed, and what the attorney must verify.
Preserve confidentiality: keep the matter's authorities and any work-product framing out of reusable templates.

Workflow

Confirm inputs. Verify the citation list (or the draft's authorities) is provided with full citations. Note which citator reports the user supplied and whether a verification connector is available. Request anything missing.
Extract the authority list. List every authority cited in the source — cases, statutes, regulations, rules, and secondary sources — with its full citation exactly as it appears and the proposition it is cited for. Flag any citation that is incomplete or malformed as [CONFIRM: citation] (and route form issues to citation-integrity-check).
State the limitation up front. Open the working notes with the explicit statement that no treatment status in this output is asserted from model memory, and that every "current" status rests on a provided citator source or an attorney's own verification.
Plan verification for each authority. For each authority, record the verification path: which supplied citator report covers it; whether a connector (e.g., CourtListener for federal case law) can confirm existence and retrieve later history; and what the attorney will still need to run in a full citator service. Note where no verification source is available at all.
Record provided citator signals. Where the user supplied a citator report, record for each authority: the report's treatment flag or signal (as the report labels it), the citing references it lists, and the report's date — each attributed to the report. Do not translate a vendor's caution flag into a legal conclusion.
Record connector findings. Where a connector was used, record that the authority exists (or that it could not be located — a prominent flag), its retrieved metadata, and any later citing documents the connector surfaced, each attributed to the connector. Do not infer good-law status from the presence or absence of citing documents.
Note high-risk authorities. Flag authorities that warrant heightened verification: old cases, cases in fast-moving or recently-legislated areas, cases central to the argument, and any authority relied on for a proposition broader than its holding (route the proposition-scope question to authority-synthesis).
Classify each authority's verification status. Assign one of: Verified current (attorney-confirmed) — only where an attorney has confirmed; Citator report provided — status per an attributed report, attorney to confirm; Connector-existence-only — existence confirmed, treatment not; Unverified — no treatment source, treat as [Verify current law]. Default to Unverified when in doubt.
Assemble the treatment-verification table and list, prominently, every authority that is Unverified or that a provided report flags for attention.
List attorney verification items and assemble the output. Consolidate placeholders, assemble the output in the format below, label it a draft for attorney review, and attach the checklist.

Output Format

Deliver the following, in order, labeled DRAFT — For Attorney Review — Treatment Not Confirmed From Model Knowledge:

Summary — one paragraph: how many authorities were checked, how many remain Unverified, and the explicit statement that no treatment status is asserted from model memory.
Verification Limitation Notice — the plain statement from Step 3.
Treatment-Verification Table — columns: Authority (full citation) | Cited For | Verification Source (report / connector / none) | Reported Signal (attributed) | Status Classification | Flags.
High-Risk Authorities — the authorities from Step 7 needing heightened verification, with the reason.
Unverified Authorities — every authority with no treatment source, each [Verify current law].
Verification Plan — for each authority, what the attorney must still run to confirm status.
Attorney Verification Items — every placeholder consolidated.
Assumptions — every assumption made, listed explicitly.

Use [Verify current law] on every status not resting on an attorney's own confirmation. Do not fill gaps with invented content.

Attorney Verification Checklist

[ ] Every authority in the source has been captured, with a complete, correct citation.
[ ] No treatment status in this draft was asserted from model knowledge; each status rests on a provided citator source or attorney verification.
[ ] Every authority has been run through a full citator service by the attorney or a qualified researcher.
[ ] Any authority a provided report flagged for negative or cautionary treatment has been read and assessed in context.
[ ] Authorities central to the argument have been independently confirmed as current good law.
[ ] For each authority, the proposition it supports has been checked against its actual holding, not just its good-law status.
[ ] Connector-existence-only authorities have had their treatment confirmed through a citator, not inferred from citing-document counts.
[ ] Old authorities and authorities in fast-moving or recently-amended areas have received heightened verification.
[ ] Statutes and regulations have been checked for currency against the official code, not only through case citators. [Verify current law]
[ ] Every Unverified authority has been resolved or removed before the draft is relied upon.
[ ] No date in this draft was computed; all dates are attributed to a provided report or connector.
[ ] The citation form of every authority has been separately verified (via citation-integrity-check).
[ ] All [Verify current law] and [CONFIRM: ...] placeholders have been resolved before the memo or brief is filed or relied upon.

Module: source-validation-workflow

Canonical path: skills/legal-methodology/source-validation/SKILL.md

name: Source Validation description: "Use when classifying whether cited sources, authorities, quotations, dates, and factual claims in draft legal work product are supported by materials available in the session, applying the source hierarchy in core/source-and-citation-discipline.md." practice_area: legal-methodology task_type: verification jurisdictions: [] risk_level: medium requires_attorney_review: true inputs:

"The draft legal work product to check"
"The cited sources, authorities, and underlying record"
"The claims, quotations, and dates to validate"

outputs:

"Claim table classifying support status for factual and legal claims"
"Source reference table"
"Unsupported claims and contradictions list"

related_skills:

skills/legal-methodology/red-team-verifier/SKILL.md
skills/legal-methodology/citation-integrity-check/SKILL.md
skills/legal-methodology/assumption-audit/SKILL.md
skills/legal-research/legal-research-memo/SKILL.md

tags:

legal-methodology
source-validation
citation-check
verification

Source Validation

Purpose

This source-validation check now uses a claim-by-claim support taxonomy. For each material factual or legal claim, classify support as source-supported, source-mentioned but insufficient, unsupported, contradicted by source, legal authority required, or attorney judgment required. These classifications organize attorney review; they do not independently verify current law or certify that a legal conclusion is correct.

Apply a systematic method for checking that every cited source, authority, quotation, date, and factual claim in a draft legal work product actually exists and says what is claimed. The skill enumerates every citation and claim in the draft, classifies each by source tier using the hierarchy in core/source-and-citation-discipline.md, and marks each as verified, unverified, or unverifiable. It produces a source-by-source validation table and a list of unresolved items that must be checked before the draft is relied upon.

This skill does not and cannot substitute for independent legal research. It does not confirm whether an authority is current, controlling, or correctly interpreted. Model background knowledge is never treated as a source, even when it appears to confirm a citation. Verification means tracing a claim to a user-provided document or a source independently retrieved and confirmed in the current session — nothing else qualifies.

Use When

A draft legal work product contains citations, quotations, authority references, or factual claims that need to be checked before the draft is finalized or passed to attorney review.
A user asks to "verify the sources," "check the citations," "confirm these quotes are real," or "validate the authority in this draft."
The Red-Team Verifier (or any review process) has flagged specific citations or claims as unverified and a dedicated source check is needed.
A legal research memo, brief section, or analysis contains cases, statutes, regulations, or secondary sources that were gathered during research and need to be confirmed before attorney reliance.
A draft quotes from a source document (contract, filing, policy, statute) and the quotations need to be checked against the provided text.
The user wants a complete audit trail showing which sources in a draft are confirmed and which require additional verification.

Choosing among the verification skills:

Use this skill for claim-by-claim classification of whether each claim in a draft is supported by the available sources.
Use red-team-verifier instead for the full adversarial pass over an entire deliverable — the superset that also covers reasoning, framing, jurisdiction, and completeness.
Use hallucination-red-team instead for a targeted hunt for fabricated or unsupported content in a draft.
Use citation-integrity-check instead when the review is limited to citations and quotations.

Required Inputs

The complete draft to be validated (uploaded or pasted). Do not validate from a summary or paraphrase.
The source documents or authorities that are supposed to support the claims in the draft — for example, the contract the draft reviews, the cases the memo cites, or the regulatory text the analysis references. These are the materials against which quotations and propositions are checked.
Optional: the skill or workflow that produced the draft — this informs which sourcing standards apply.
Optional: a list of specific citations or claims flagged for priority review.

If the draft is not provided, stop and request it. If source documents are not provided, the validation will be limited to classifying citations by tier and marking unsupported claims as unverifiable — note this limitation prominently.

Do Not Use When

The task is to conduct new legal research and find sources for unsupported propositions — use legal-research-memo or a research skill; this skill validates existing sources, it does not find new ones.
The task is to conduct a broader structural review of draft quality — use the Red-Team Verifier for that purpose; use this skill for focused source and citation checking.
The draft contains no citations, quotations, or factual claims traceable to external sources — there is nothing to validate.
The user needs an attorney to assess whether an authority supports a legal argument — that is a substantive judgment beyond the scope of this skill; flag those items as attorney verification items.

Legal Safety Rules

Produce draft legal work product for attorney review. This is not legal advice.
Model background knowledge is never a verification source. A citation that appears consistent with what the model knows about a case or statute is still classified as unverified unless it is present in a user-provided document or independently retrieved and confirmed in this session.
Do not assert that a citation is valid or that a quotation is accurate based on model background knowledge, training data, or general familiarity with legal sources.
Do not invent, reconstruct, or complete a partial citation. If a citation is incomplete, mark it as incomplete and list it as an unresolved item.
Do not assert that a case, statute, or regulation does not exist — only that it could not be verified within this session. Absence of verification is not confirmation of fabrication; that determination requires independent research.
Do not assess whether an authority is controlling, persuasive, current, or correctly applied — those are attorney functions. This skill checks existence and textual accuracy, not legal significance.
Do not place client-sensitive facts from the draft into any reusable output beyond this validation report.
Label the validation report as draft legal work product for attorney review.

Workflow

Confirm inputs. Verify that you have the complete draft text. Note whether source documents have been provided and, if not, record the resulting limitation: without source documents, only citation-tier classification is possible — quotation checking and claim-source tracing require the underlying documents.

Enumerate every citation and claim. Use skills/legal-research/references/citation-type-taxonomy.md to classify each citation's form (full case citation, short-form, supra, Id., statutory, regulation, law-journal, or unknown/partial) so that short forms, supras, and Id. citations are correctly matched to their antecedent before verification. Also assign each citation a verification tier using skills/legal-methodology/references/citation-confidence-tiers.md alongside the source-tier classification in step 3 below, so the report distinguishes citation form from verification status. Read through the draft systematically and build a complete inventory of:
- Case citations: case name, citation, reporter, year, court.
- Statutory and regulatory citations: instrument name, section, version or effective date.
- Secondary source citations: treatise, article, report, guidance document.
- Quotations: every passage presented as a direct quotation from any source.
- Factual claims: every claim of fact that depends on a source document, client representation, or external authority — as distinct from analytical inference.
- Dates: every specific date asserted in the draft, including effective dates, filing dates, events, and deadlines.

Assign each item a unique number for tracking. If the draft is long, work section by section and flag the section with each entry.

For every material factual or legal claim, also assign one source-support status: Source-supported, Source-mentioned but insufficient, Unsupported, Contradicted by source, Legal authority required, or Attorney judgment required. Use "source-supported" only when an available source directly supports the claim as stated, subject to attorney review.

Classify each item by source tier. Apply the hierarchy from core/source-and-citation-discipline.md:
- Tier 1 — User-provided document. The claim or quotation is traceable to a document the user supplied in this session. This is the highest-trust tier; verify by checking the text of the provided document.
- Tier 2 — Independently researched and verified. The claim was located through a research step and confirmed to exist and to say what is claimed, through a source retrieved in this session. Identify the verification step.
- Tier 3 — Model background knowledge. The claim appears consistent with model training knowledge but has not been confirmed against a Tier 1 or Tier 2 source. Tier 3 is never a valid verification source.

Verify Tier 1 items. For every claim or quotation classified as Tier 1:
- Locate the specific text in the provided document.
- Confirm that the quotation matches the source text exactly.
- Confirm that the section reference, page, or paragraph is accurate.
- Confirm that the claim accurately represents what the source says (not overstated, not understated).
- Mark as: Verified — Tier 1, with a note identifying the source document and location.
- If the claim cannot be located in the provided document, reclassify as Unverified and note the discrepancy.

Assess Tier 2 items. For every claim classified as Tier 2 (independently researched):
- Identify whether the research step and verification are documented in the draft or the underlying workflow.
- If documented and confirmable in this session, mark as Verified — Tier 2, with a note identifying the verification step.
- If not documented or not confirmable in this session, mark as Unverified — requires independent confirmation and list as an unresolved item.

Flag all Tier 3 items. For every claim that relies only on model background knowledge:
- Mark as Unverified — Tier 3 (model knowledge only); must be independently confirmed before reliance.
- Do not treat apparent consistency with model knowledge as verification.
- List every Tier 3 item as an unresolved item requiring independent research.

Identify unverifiable items. Flag items that cannot be verified within this session and whose verification requires access to external legal databases, official sources, or documents not provided:
- Case citations where the case text was not provided.
- Statutory text where the current version was not provided.
- Regulatory text where the regulation was not provided.
- Quotations from sources not provided. Mark each as Unverifiable in this session — must be confirmed by attorney or researcher.

Check for incomplete and malformed citations. For every citation in the inventory, check:
- Is the citation complete? (Case: name, reporter, volume, page, court, year. Statute: instrument, section, version. Regulation: title, part, section, version.)
- Are there missing elements (for example, a case name without a reporter cite, or a statute without a section number)?
- Are there obvious formatting errors (for example, mismatched parenthetical dates or transposed numbers)? Mark incomplete citations as defects and list them as unresolved items.

Check date claims. For every date in the inventory:
- Identify whether it is a user-provided date, a date from a provided document, or an asserted date with no source.
- Flag any date that appears to have been computed or derived (rather than quoted from a source) as [deadline verification required].
- Flag any effective date or filing date that has not been confirmed against the applicable source.

Compile unresolved items. Build a consolidated list of all items marked Unverified, Unverifiable, or defective. For each, state: what the item is, why it is unresolved, and what step is needed to resolve it.

Assemble the validation report using the output format below. Label it as a draft for attorney review.

Output Format

Deliver a Source Validation Report with the following sections:

Report Header — Draft title or description; date of validation; count of items (total | verified | unverified | unverifiable | defective); overall status (Cleared for attorney review / Unresolved items present — see below).
Scope and Limitations — Whether source documents were provided; any limitations on what could be verified in this session; explicit statement that model background knowledge was not used as a verification source.
Claim Table — One row per material claim:

#	Claim	Claim type	Source cited or needed	Support status	Revision needed

Source Reference Table — One row per source:

Source	Source type	Provided by	Claims supported	Limits / notes

Source Validation Table — One row per enumerated item:

#	Item Type	Claim / Citation as Stated	Asserted Source	Source Tier	Status	Notes

Status values: Verified — Tier 1 | Verified — Tier 2 | Unverified | Unverifiable in this session | Defective (incomplete / malformed)

Unsupported Claims List — Every unsupported or insufficiently supported claim, with required action.
Contradictions List — Every claim contradicted by available source material.
Recommended Revisions — Safer wording, placeholders, or deletion recommendations for unresolved claims.
Unresolved Items List — A numbered list of all unverified, unverifiable, and defective items, each stating: what needs to be checked, by what method, and who is responsible (attorney / legal researcher / client).
Date and Deadline Flags — A separate list of all date claims flagged [deadline verification required] or lacking a confirmed source.
Overall Status Statement — Either: "No unresolved items. Draft cleared for attorney review, subject to attorney verification checklist." Or: "Unresolved items present. Draft must not be relied upon until all unresolved items are confirmed."

Label the report: Draft legal work product for attorney review. Not legal advice.

Attorney Verification Checklist

[ ] Every Tier 1 (user-provided document) verification has been confirmed: the attorney has checked that the quotation and claim match the source document.
[ ] Every Tier 2 (independently researched) item has been confirmed: the research step is documented and the authority has been independently verified.
[ ] Every item marked Unverified or Unverifiable has been independently checked against an authoritative source by the attorney or a qualified researcher.
[ ] Every case citation has been confirmed to exist, to be correctly cited, and to stand for the proposition the draft attributes to it.
[ ] Every statute and regulation citation has been confirmed in the current version applicable to the matter [deadline verification required].
[ ] Every quotation has been checked against its source text and confirmed to be accurate and in context.
[ ] Every defective (incomplete or malformed) citation has been corrected or removed.
[ ] No date in the draft was computed by the agent; all dates are confirmed against source documents or attorney knowledge [deadline verification required].
[ ] No authority has been accepted as verified on the basis of model background knowledge alone.
[ ] All items on the Unresolved Items List have been resolved and documented.
[ ] The validation report and any corrections are retained as part of the matter file.
[ ] The draft is appropriately labeled as draft legal work product for attorney review, not legal advice.

Purpose

Produce a structured, attorney-ready legal research memo in response to a discrete legal question. The memo organizes analysis using IRAC-style discipline — Question Presented, Brief Answer, Facts, Assumptions, Discussion/Analysis, Conclusion — and maintains a strict separation between facts, assumptions, applicable law, analysis, and open verification items. It produces draft legal work product for attorney review only. It is not legal advice and does not substitute for attorney judgment.

The most important discipline this skill enforces: no legal authority — no case, statute, regulation, rule, secondary source, or quotation — may be stated as if it exists unless it comes from a user-provided document or has been independently researched and verified through a reliable source. Every asserted authority must be checkable. Unverified authority must be marked with an explicit [CONFIRM: ...] placeholder and placed in the attorney verification section.

Use When

A user asks to "research this issue," "write a research memo," "what is the law on X," or "can you analyze whether Y is legal."
A lawyer or legal team needs a first-pass research memo before attorney analysis.
The user needs to organize known authorities and facts into a structured memo before a client call, brief, or negotiation.
A question of law or mixed fact-and-law has been identified and needs structured written analysis.
The user wants to document legal assumptions underlying a transaction, filing, or decision.
A gap analysis is needed: what authorities support a position, and what do not.

Required Inputs

The legal question(s) — stated with specificity. Vague questions must be refined before proceeding; do not broaden or narrow the question without user confirmation.
Jurisdiction and governing law — the applicable jurisdiction(s) and the body of law (federal, state, contractual, regulatory). If unknown, flag as [CONFIRM: jurisdiction] and note that the analysis cannot be completed without it.
Relevant facts — the facts from which the legal question arises. Do not reconstruct or invent facts; use only facts the user has provided.
Procedural posture — if applicable (e.g., pre-litigation, pending motion, transactional, regulatory inquiry).
Relevant date — the date on which the legal question turns (e.g., the date of the alleged breach, the filing date, the effective date of a statute).
Optional: any legal authorities, cases, statutes, or secondary sources the user has already identified. These will be incorporated and verified, not assumed to be accurate.
Optional: the user's desired legal position or outcome (flags potential advocacy posture).

If the legal question, jurisdiction, or relevant facts are not provided, stop and request them. Do not begin analysis by guessing. Do not fabricate facts to make the question answerable.

Do Not Use When

The user needs a contract review (use nda-review or contract-risk-review).
The user needs a section of a litigation brief drafted (use brief-section-drafter).
The user needs a regulatory compliance checklist (use the appropriate compliance skill).
The question requires a formal legal opinion letter — those require an attorney and carry professional responsibility implications this skill cannot satisfy.
The user is asking for real-time legal advice in an ongoing matter requiring immediate attorney judgment.

Legal Safety Rules

Produce draft legal work product for attorney review. This is not legal advice.
Source and citation discipline. Follow core/source-and-citation-discipline.md. Never invent legal authority, citations, quotations, statutes, cases, regulations, filing deadlines, or procedural rules. Label what is a provided source, a user-provided fact, an assumption, a legal inference, or an item requiring attorney verification, and use a citation placeholder such as [Attorney to insert authority] when no source is available. Never state a case name, citation, holding, statute section, regulation, or quotation unless it is present in a user-provided document or has been retrieved from and verified through a reliable research source. Fabricated citations are a serious professional and ethical hazard.
Every legal authority cited must be either (a) provided by the user or (b) explicitly noted as requiring attorney verification if it cannot be independently confirmed in this session.
Do not paraphrase a case holding without citing the source. Do not cite a source without identifying where the specific proposition comes from.
Distinguish clearly: (1) what the facts show, (2) what you are assuming, (3) what the law provides (with source), (4) what the analysis concludes, and (5) what remains for the attorney to confirm.
Do not assume a statute or regulation is current. Flag recency as an attorney verification item unless the user has provided the current text.
Do not resolve ambiguous facts in favor of either party without flagging that assumption explicitly.
Identify any conflict of laws, choice-of-law, or multi-jurisdictional issue as an attorney verification item.
Do not place client-sensitive facts into reusable templates.
Use [CONFIRM: ...] placeholders wherever information is missing or uncertain. Never fill gaps with invented content.
Consult connectors/ when a verification path is available. When the session has access to a case-law connector — for US federal case law, connectors/courtlistener.md — resolve each [VERIFY-CITE: ...] placeholder by querying the connector and recording the verified URL in the Authorities Cited table. A verified citation still requires attorney verification of the proposition it stands for; the connector closes the "does this case exist" gap, not the "does it support this proposition" gap. When no connector is available, the placeholder remains; the discipline never relaxes.

Workflow

Confirm inputs. Verify that you have a specific legal question, identified jurisdiction(s), relevant facts, and a relevant date. If anything is missing, request it before proceeding. Refine vague questions by asking clarifying questions — do not assume scope.

Restate the Question(s) Presented. Frame each legal question precisely: who did what, under what legal framework, in what jurisdiction, at what time. Limit each question to one discrete legal issue. If the user's question spans multiple issues, break it into separate Questions Presented.

Draft the Brief Answer(s). Provide a one-to-three sentence direct answer to each question: yes, no, probably, or uncertain — with a one-sentence reason. Do not elaborate here; save analysis for the Discussion section. If the answer depends entirely on unverified facts or law, say so explicitly.

State the Facts. Set out the legally relevant facts as provided by the user. Do not embellish, infer, or add facts not in the user's account. Note any factual gaps that are legally material. Label inferences as inferences, not facts.

State Assumptions. List every assumption being made — about facts, about the applicable legal standard, about the procedural posture, about the status of authorities. Each assumption is a potential attorney verification item.

Conduct and document the Discussion/Analysis. Apply an IRAC structure for each issue:
- Issue — restate the specific sub-question.
- Rule — state the applicable legal rule or standard, citing the source precisely. If the rule comes from a case, identify the court, date, and the specific holding. If from a statute or regulation, identify the section and version. If no verified rule is available, use [CONFIRM: rule — no verified authority in this session].
- Application — apply the rule to the facts. Flag where the facts are disputed, unclear, or assumed. Do not advocate silently; if analysis favors one side, note that.
- Conclusion — state the tentative conclusion for the sub-issue, with confidence level (e.g., "likely," "uncertain," "turns on factual dispute").

Write the Conclusion. Summarize the overall answer to each Question Presented in two to five sentences. Identify the key legal and factual variables on which the answer depends. Do not overstate certainty.

Compile the Authorities Cited table. List every legal authority referenced in the memo — cases, statutes, regulations, secondary sources. For each, include: authority name/citation, source (user-provided, researched, or verified via connector), proposition it stands for, and a verification checkbox. For US federal case citations, attempt verification through CourtListener when the session has access to it — see connectors/courtlistener.md for the calling pattern, in-scope corpus, and fallback behavior. When a citation is verified, record the connector URL in the table and replace the [VERIFY-CITE: ...] placeholder with [ATTORNEY TO CONFIRM: proposition supported by the cited case]. When no connector is available, or the citation falls outside the connector's in-scope corpus, retain the placeholder unchanged. See templates/legal-research-memo.md.

List Open Items for Attorney Verification. Enumerate every factual gap, unverified authority, jurisdictional question, ambiguity, or strategic judgment that requires attorney review before the memo is relied upon.

Assemble the memo using templates/legal-research-memo.md and label it as a draft for attorney review.

Output Format

Deliver a complete memo using the structure in templates/legal-research-memo.md:

Header block — To, From, Date, Re, Privilege designation.
Question(s) Presented — numbered, each limited to one issue.
Brief Answer(s) — direct, one-to-three sentences each.
Facts — user-provided facts only, clearly labeled.
Assumptions — explicit list.
Discussion / Analysis — IRAC per issue, with source citations for every rule stated.
Conclusion — summary with qualified confidence.
Authorities Cited — table with verification-source column (user-provided, connector-verified with URL, or [VERIFY-CITE: ...]) and verification checkbox column.
Open Items / Attorney Verification — checkbox list.

Use [CONFIRM: ...] wherever a fact, authority, or conclusion is unverified. Do not omit a section because it is difficult to fill; instead, mark it with a placeholder and a note.

Attorney Verification Checklist

[ ] The legal question(s) are accurately and completely stated as the client intends them.
[ ] Jurisdiction and governing law are correctly identified and appropriate for the matter.
[ ] All facts stated in the memo are accurate and come from the client or verified sources — no facts have been invented or inferred without flagging.
[ ] All assumptions are identified and their legal materiality has been assessed.
[ ] Every case cited exists, the citation is accurate, and the holding attributed to it is correct.
[ ] Every statute and regulation cited is current, in the correct version, and the section referenced says what the memo claims.
[ ] No authority has been cited that was not in a user-provided document or independently verified in this session (including via a connectors/ source, where applicable).
[ ] The rule/holding has not been overstated, paraphrased inaccurately, or taken out of context.
[ ] Adverse authority has been identified and addressed, not omitted.
[ ] Any conflict-of-laws or choice-of-law issue has been identified and resolved or flagged.
[ ] The analysis is presented from a neutral analytical posture, or the advocacy posture is explicitly noted.
[ ] Confidence levels in the conclusion are appropriate given the state of the law and facts.
[ ] All open items and [CONFIRM: ...] placeholders have been resolved before the memo is relied upon.
[ ] Privilege and confidentiality designations are appropriate for how the memo will be circulated.

Core raw SKILL.md

---
name: Legal Research Memo
description: "Use when producing a structured legal research memo in response to a specific legal question, organizing analysis using IRAC discipline (Question Presented, Brief Answer, Facts, Assumptions, Discussion/Analysis, Conclusion) with explicit sourcing requirements and attorney verification checkpoints."
practice_area: legal-research
task_type: research
jurisdictions: []
risk_level: medium
requires_attorney_review: true
inputs:
- "The specific legal question presented"
- "The known facts"
- "The applicable jurisdiction"
- "Any authority already provided"
outputs:
- "Structured legal research memo using IRAC discipline for attorney review"
related_skills:
- skills/litigation/brief-section-drafter/SKILL.md
- skills/legal-research/negative-treatment-check/SKILL.md
- skills/legal-methodology/statutory-interpretation/SKILL.md
- skills/legal-methodology/source-validation/SKILL.md
tags:
- legal-research
- research-memo
- irac
- legal-analysis
- citations
---

# Legal Research Memo

## Purpose

The most important discipline this skill enforces: **no legal authority — no case, statute, regulation, rule, secondary source, or quotation — may be stated as if it exists unless it comes from a user-provided document or has been independently researched and verified through a reliable source.** Every asserted authority must be checkable. Unverified authority must be marked with an explicit `[CONFIRM: ...]` placeholder and placed in the attorney verification section.

## Use When

- A user asks to "research this issue," "write a research memo," "what is the law on X," or "can you analyze whether Y is legal."
- A lawyer or legal team needs a first-pass research memo before attorney analysis.
- The user needs to organize known authorities and facts into a structured memo before a client call, brief, or negotiation.
- A question of law or mixed fact-and-law has been identified and needs structured written analysis.
- The user wants to document legal assumptions underlying a transaction, filing, or decision.
- A gap analysis is needed: what authorities support a position, and what do not.

## Required Inputs

- **The legal question(s)** — stated with specificity. Vague questions must be refined before proceeding; do not broaden or narrow the question without user confirmation.
- **Jurisdiction and governing law** — the applicable jurisdiction(s) and the body of law (federal, state, contractual, regulatory). If unknown, flag as `[CONFIRM: jurisdiction]` and note that the analysis cannot be completed without it.
- **Relevant facts** — the facts from which the legal question arises. Do not reconstruct or invent facts; use only facts the user has provided.
- **Procedural posture** — if applicable (e.g., pre-litigation, pending motion, transactional, regulatory inquiry).
- **Relevant date** — the date on which the legal question turns (e.g., the date of the alleged breach, the filing date, the effective date of a statute).
- Optional: any legal authorities, cases, statutes, or secondary sources the user has already identified. These will be incorporated and verified, not assumed to be accurate.
- Optional: the user's desired legal position or outcome (flags potential advocacy posture).

If the legal question, jurisdiction, or relevant facts are not provided, stop and request them. Do not begin analysis by guessing. Do not fabricate facts to make the question answerable.

## Do Not Use When

- The user needs a contract review (use `nda-review` or `contract-risk-review`).
- The user needs a section of a litigation brief drafted (use `brief-section-drafter`).
- The user needs a regulatory compliance checklist (use the appropriate compliance skill).
- The question requires a formal legal opinion letter — those require an attorney and carry professional responsibility implications this skill cannot satisfy.
- The user is asking for real-time legal advice in an ongoing matter requiring immediate attorney judgment.

## Legal Safety Rules

- Produce draft legal work product for attorney review. This is not legal advice.
- **Source and citation discipline.** Follow `core/source-and-citation-discipline.md`. Never invent legal authority, citations, quotations, statutes, cases, regulations, filing deadlines, or procedural rules. Label what is a provided source, a user-provided fact, an assumption, a legal inference, or an item requiring attorney verification, and use a citation placeholder such as `[Attorney to insert authority]` when no source is available. Never state a case name, citation, holding, statute section, regulation, or quotation unless it is present in a user-provided document or has been retrieved from and verified through a reliable research source. Fabricated citations are a serious professional and ethical hazard.
- Every legal authority cited must be either (a) provided by the user or (b) explicitly noted as requiring attorney verification if it cannot be independently confirmed in this session.
- Do not paraphrase a case holding without citing the source. Do not cite a source without identifying where the specific proposition comes from.
- Distinguish clearly: (1) what the facts show, (2) what you are assuming, (3) what the law provides (with source), (4) what the analysis concludes, and (5) what remains for the attorney to confirm.
- Do not assume a statute or regulation is current. Flag recency as an attorney verification item unless the user has provided the current text.
- Do not resolve ambiguous facts in favor of either party without flagging that assumption explicitly.
- Identify any conflict of laws, choice-of-law, or multi-jurisdictional issue as an attorney verification item.
- Do not place client-sensitive facts into reusable templates.
- Use `[CONFIRM: ...]` placeholders wherever information is missing or uncertain. Never fill gaps with invented content.
- **Consult `connectors/` when a verification path is available.** When the session has access to a case-law connector — for US federal case law, `connectors/courtlistener.md` — resolve each `[VERIFY-CITE: ...]` placeholder by querying the connector and recording the verified URL in the Authorities Cited table. A verified citation still requires attorney verification of the proposition it stands for; the connector closes the "does this case exist" gap, not the "does it support this proposition" gap. When no connector is available, the placeholder remains; the discipline never relaxes.

## Workflow

1. **Confirm inputs.** Verify that you have a specific legal question, identified jurisdiction(s), relevant facts, and a relevant date. If anything is missing, request it before proceeding. Refine vague questions by asking clarifying questions — do not assume scope.

2. **Restate the Question(s) Presented.** Frame each legal question precisely: who did what, under what legal framework, in what jurisdiction, at what time. Limit each question to one discrete legal issue. If the user's question spans multiple issues, break it into separate Questions Presented.

3. **Draft the Brief Answer(s).** Provide a one-to-three sentence direct answer to each question: yes, no, probably, or uncertain — with a one-sentence reason. Do not elaborate here; save analysis for the Discussion section. If the answer depends entirely on unverified facts or law, say so explicitly.

4. **State the Facts.** Set out the legally relevant facts as provided by the user. Do not embellish, infer, or add facts not in the user's account. Note any factual gaps that are legally material. Label inferences as inferences, not facts.

5. **State Assumptions.** List every assumption being made — about facts, about the applicable legal standard, about the procedural posture, about the status of authorities. Each assumption is a potential attorney verification item.

6. **Conduct and document the Discussion/Analysis.** Apply an IRAC structure for each issue:
- **Issue** — restate the specific sub-question.
- **Rule** — state the applicable legal rule or standard, citing the source precisely. If the rule comes from a case, identify the court, date, and the specific holding. If from a statute or regulation, identify the section and version. If no verified rule is available, use `[CONFIRM: rule — no verified authority in this session]`.
- **Application** — apply the rule to the facts. Flag where the facts are disputed, unclear, or assumed. Do not advocate silently; if analysis favors one side, note that.
- **Conclusion** — state the tentative conclusion for the sub-issue, with confidence level (e.g., "likely," "uncertain," "turns on factual dispute").

7. **Write the Conclusion.** Summarize the overall answer to each Question Presented in two to five sentences. Identify the key legal and factual variables on which the answer depends. Do not overstate certainty.

8. **Compile the Authorities Cited table.** List every legal authority referenced in the memo — cases, statutes, regulations, secondary sources. For each, include: authority name/citation, source (user-provided, researched, or verified via connector), proposition it stands for, and a verification checkbox. For US federal case citations, attempt verification through CourtListener when the session has access to it — see `connectors/courtlistener.md` for the calling pattern, in-scope corpus, and fallback behavior. When a citation is verified, record the connector URL in the table and replace the `[VERIFY-CITE: ...]` placeholder with `[ATTORNEY TO CONFIRM: proposition supported by the cited case]`. When no connector is available, or the citation falls outside the connector's in-scope corpus, retain the placeholder unchanged. See `templates/legal-research-memo.md`.

9. **List Open Items for Attorney Verification.** Enumerate every factual gap, unverified authority, jurisdictional question, ambiguity, or strategic judgment that requires attorney review before the memo is relied upon.

10. **Assemble the memo** using `templates/legal-research-memo.md` and label it as a draft for attorney review.

## Output Format

Deliver a complete memo using the structure in `templates/legal-research-memo.md`:

1. **Header block** — To, From, Date, Re, Privilege designation.
2. **Question(s) Presented** — numbered, each limited to one issue.
3. **Brief Answer(s)** — direct, one-to-three sentences each.
4. **Facts** — user-provided facts only, clearly labeled.
5. **Assumptions** — explicit list.
6. **Discussion / Analysis** — IRAC per issue, with source citations for every rule stated.
7. **Conclusion** — summary with qualified confidence.
8. **Authorities Cited** — table with verification-source column (user-provided, connector-verified with URL, or `[VERIFY-CITE: ...]`) and verification checkbox column.
9. **Open Items / Attorney Verification** — checkbox list.

Use `[CONFIRM: ...]` wherever a fact, authority, or conclusion is unverified. Do not omit a section because it is difficult to fill; instead, mark it with a placeholder and a note.

## Attorney Verification Checklist

- [ ] The legal question(s) are accurately and completely stated as the client intends them.
- [ ] Jurisdiction and governing law are correctly identified and appropriate for the matter.
- [ ] All facts stated in the memo are accurate and come from the client or verified sources — no facts have been invented or inferred without flagging.
- [ ] All assumptions are identified and their legal materiality has been assessed.
- [ ] Every case cited exists, the citation is accurate, and the holding attributed to it is correct.
- [ ] Every statute and regulation cited is current, in the correct version, and the section referenced says what the memo claims.
- [ ] No authority has been cited that was not in a user-provided document or independently verified in this session (including via a `connectors/` source, where applicable).
- [ ] The rule/holding has not been overstated, paraphrased inaccurately, or taken out of context.
- [ ] Adverse authority has been identified and addressed, not omitted.
- [ ] Any conflict-of-laws or choice-of-law issue has been identified and resolved or flagged.
- [ ] The analysis is presented from a neutral analytical posture, or the advocacy posture is explicitly noted.
- [ ] Confidence levels in the conclusion are appropriate given the state of the law and facts.
- [ ] All open items and `[CONFIRM: ...]` placeholders have been resolved before the memo is relied upon.
- [ ] Privilege and confidentiality designations are appropriate for how the memo will be circulated.

AgentCounsel selective execution package for Legal Research Memo.

The core skill always applies. Choose one declared execution mode and evaluate each module's machine-readable activation object exactly. Missing activation inputs fail closed: do not load every conditional module. A module without activation is not selected unless it is required or explicitly requested. Use only the selected resources, preserve why each resource was selected, and keep every parent safety rule active.

This portable package contains every possible resource. For true prompt-size reduction, use the MCP get_skill_context tool or another client that sends only the selected bundle.

=== BEGIN AGENTCOUNSEL CORE SKILL ===
---
name: Legal Research Memo
description: "Use when producing a structured legal research memo in response to a specific legal question, organizing analysis using IRAC discipline (Question Presented, Brief Answer, Facts, Assumptions, Discussion/Analysis, Conclusion) with explicit sourcing requirements and attorney verification checkpoints."
practice_area: legal-research
task_type: research
jurisdictions: []
risk_level: medium
requires_attorney_review: true
inputs:
- "The specific legal question presented"
- "The known facts"
- "The applicable jurisdiction"
- "Any authority already provided"
outputs:
- "Structured legal research memo using IRAC discipline for attorney review"
related_skills:
- skills/litigation/brief-section-drafter/SKILL.md
- skills/legal-research/negative-treatment-check/SKILL.md
- skills/legal-methodology/statutory-interpretation/SKILL.md
- skills/legal-methodology/source-validation/SKILL.md
tags:
- legal-research
- research-memo
- irac
- legal-analysis
- citations
---

# Legal Research Memo

## Purpose

## Use When

## Required Inputs

If the legal question, jurisdiction, or relevant facts are not provided, stop and request them. Do not begin analysis by guessing. Do not fabricate facts to make the question answerable.

## Do Not Use When

## Legal Safety Rules

## Workflow

10. **Assemble the memo** using `templates/legal-research-memo.md` and label it as a draft for attorney review.

## Output Format

Deliver a complete memo using the structure in `templates/legal-research-memo.md`:

Use `[CONFIRM: ...]` wherever a fact, authority, or conclusion is unverified. Do not omit a section because it is difficult to fill; instead, mark it with a placeholder and a note.

## Attorney Verification Checklist

=== END AGENTCOUNSEL CORE SKILL ===

=== BEGIN AGENTCOUNSEL SPEC ===
Canonical path: skills/legal-research/legal-research-memo/SPEC.json
{
"schema_version": "2.0",
"skill_id": "legal-research/legal-research-memo",
"gates": {
"custom": [
{
"id": "specific-question-required",
"condition": "The legal question is vague, compound, or materially underspecified.",
"action": "stop-and-ask",
"reason": "The research scope must be confirmed before authorities are gathered or conclusions are drafted."
},
{
"id": "verified-authority-required",
"condition": "A legal proposition would be stated without user-provided or independently verified authority.",
"action": "require-attorney-confirmation",
"reason": "Unverified legal authority must remain a visible placeholder and may not be presented as existing or current law."
},
{
"id": "formal-opinion-boundary",
"condition": "The requested work product is a formal legal opinion letter or an immediate client-facing legal conclusion.",
"action": "stop-and-escalate",
"reason": "Formal opinions and real-time legal advice require direct attorney authorship and professional-responsibility review."
}
]
},
"execution_modes": [
{
"id": "quick-triage",
"enabled": true,
"purpose": "Refine the question presented, identify jurisdiction and factual gaps, propose a research plan, and list authority types to verify without drafting a substantive conclusion.",
"output_detail": "minimal",
"quality_checks": [
"attorney-review-gate",
"jurisdiction-deadline-gates"
]
},
{
"id": "standard",
"enabled": true,
"purpose": "Produce the complete attorney-review research memo with verified authority, explicit assumptions, qualified analysis, and open verification items.",
"output_detail": "standard",
"quality_checks": [
"attorney-review-gate",
"citation-integrity-check",
"source-validation-check",
"assumption-audit",
"legal-prose-polish"
]
},
{
"id": "deep-review",
"enabled": true,
"purpose": "Run the standard memo plus adverse-authority, currentness, negative-treatment, proposition-support, and citation-integrity passes for each material proposition.",
"output_detail": "expanded",
"quality_checks": [
"attorney-review-gate",
"citation-integrity-check",
"source-validation-check",
"assumption-audit",
"hallucination-red-team",
"output-format-compliance-check",
"privilege-confidentiality-check",
"legal-prose-polish"
]
}
],
"input_schema": [
{
"id": "specific-legal-question-presented",
"label": "Specific legal question presented",
"type": "text",
"required": true,
"description": "A discrete legal question with confirmed scope.",
"source_requirement": "user-provided-and-confirmed",
"may_infer": false,
"sensitive": false
},
{
"id": "known-facts",
"label": "Known facts",
"type": "object",
"required": true,
"description": "The legally material facts, each attributable to the user or a provided source.",
"source_requirement": "user-provided-or-provided-document",
"may_infer": false,
"sensitive": true
},
{
"id": "applicable-jurisdiction",
"label": "Applicable jurisdiction",
"type": "jurisdiction",
"required": true,
"description": "The jurisdiction and body of law governing the question.",
"source_requirement": "user-provided-and-attorney-confirmed",
"may_infer": false,
"sensitive": false
},
{
"id": "authority-already-provided",
"label": "Authority already provided",
"type": "document-set",
"required": false,
"description": "Cases, statutes, regulations, rules, or secondary sources already supplied for verification and use.",
"source_requirement": "provided-document",
"may_infer": false,
"sensitive": false
}
],
"output_schema": [
{
"id": "structured-legal-research-memo-using-irac-discipline-for-attorney-review",
"label": "Structured legal research memo using IRAC discipline for attorney review",
"type": "memo",
"required": true,
"description": "Question presented, brief answer, facts, assumptions, issue-by-issue analysis, conclusion, authorities, and open attorney-verification items.",
"attorney_review_required": true
},
{
"id": "claim-authority-map",
"label": "Claim-to-authority map",
"type": "table",
"required": true,
"description": "Every material legal proposition mapped to the exact authority and locator that supports it, including verification status and contrary authority.",
"attorney_review_required": true
}
],
"evidence_schema": {
"required_fields": [
"support_text",
"authority_status",
"proposition_supported"
],
"fields": [
{
"id": "authority_status",
"type": "enum",
"description": "Whether the authority is user-provided, connector-verified, independently verified, secondary-only, or unresolved."
},
{
"id": "proposition_supported",
"type": "text",
"description": "The precise proposition for which the source is cited."
},
{
"id": "negative_treatment_status",
"type": "enum",
"description": "The recorded treatment status when a negative-treatment check is performed."
}
]
},
"modules": [
{
"id": "legal-research-memo-template",
"kind": "template",
"path": "skills/legal-research/legal-research-memo/templates/legal-research-memo.md",
"required": true,
"load_when": "standard or deep-review mode"
},
{
"id": "citation-type-taxonomy",
"kind": "reference",
"path": "skills/legal-research/references/citation-type-taxonomy.md",
"required": false,
"load_when": "the memo uses multiple authority types or citation forms"
},
{
"id": "courtlistener-connector-guide",
"kind": "connector",
"path": "connectors/courtlistener.md",
"required": false,
"load_when": "US federal case existence or citation verification is available through CourtListener"
},
{
"id": "negative-treatment-workflow",
"kind": "workflow-module",
"path": "skills/legal-research/negative-treatment-check/SKILL.md",
"required": false,
"load_when": "deep-review mode or when a material case controls the analysis"
},
{
"id": "source-validation-workflow",
"kind": "quality-check",
"path": "skills/legal-methodology/source-validation/SKILL.md",
"required": true,
"load_when": "standard or deep-review mode"
}
]
}

=== END AGENTCOUNSEL SPEC ===

=== BEGIN AGENTCOUNSEL MODULE: legal-research-memo-template ===
Kind: template
Canonical path: skills/legal-research/legal-research-memo/templates/legal-research-memo.md
Load when: standard or deep-review mode

# Legal Research Memo

> Draft legal work product for attorney review. Not legal advice.
> Do not paste client-sensitive facts into a reusable copy of this template.

---

**PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT**

| | |
|---|---|
| **To:** | [Supervising Attorney / Matter Team] |
| **From:** | [Preparer Name / AgentCounsel workflow] |
| **Date:** | [Date of memo] |
| **Re:** | [Brief description of legal question and matter name or number] |
| **Matter:** | [Matter name / number] |
| **Jurisdiction:** | [CONFIRM: jurisdiction and governing law] |
| **Relevant Date:** | [Date on which the legal question turns — e.g., date of alleged breach, filing date] |

---

## Question(s) Presented

*State each legal question precisely: who did what, under which legal framework, in which jurisdiction, and at what time. Limit each question to one discrete legal issue.*

1. [Question 1: e.g., "Under [governing law], did [party] [action] so as to [legal consequence] as of [date]?"]
2. [Question 2, if applicable]
3. [Add as needed]

---

## Brief Answer(s)

1. **[Yes / No / Probably / Uncertain].** [One-to-two sentence reason. E.g., "Probably yes, because [key rule] appears to apply to [key fact], though [key variable] remains unconfirmed and is a material condition of this answer."]
2. [Answer to Question 2, if applicable]
3. [Add as needed]

---

## Facts

- [Fact 1 — as provided]
- [Fact 2 — as provided]
- [Inference: [describe inference] — **noted as inference, not established fact**]
- [CONFIRM: [describe factual gap] — legally material; obtain before relying on this memo]
- [Add as needed]

---

## Assumptions

- [Assumption 1: e.g., "The agreement referenced is governed by [state] law as stated in the governing-law clause; this has not been independently verified."]
- [Assumption 2: e.g., "The relevant date for purposes of the statute of limitations analysis is [date]; attorney should confirm which event triggers the limitations period."]
- [Assumption 3: e.g., "No amendments to the referenced statute have been enacted between [date] and the date of this memo; attorney should confirm current version."]
- [Add as needed]

---

## Discussion / Analysis

---

### Issue 1: [Restate the first legal sub-question]

**Rule**

[State the applicable legal rule or standard. Identify the source precisely: case name, court, year, and specific holding; or statute/regulation with section number and version. Example: "Under [Statute], § [X], a party must [requirement]. [Citation / source]. If the authority cannot be verified in this session, write: `[CONFIRM: rule — no verified authority available for this proposition; attorney must supply and verify before relying on this analysis.]`"]

**Application**

- Element 1 ([element name]): [Does the fact pattern satisfy this element? Why or why not?]
- Element 2 ([element name]): [Analysis]
- Element 3 ([element name]): [Analysis — if uncertain: `[CONFIRM: whether [fact] satisfies this element]`]

**Conclusion on Issue 1**

---

### Issue 2: [Restate the second legal sub-question, if applicable]

**Rule**

[Rule with citation / source, or `[CONFIRM]` placeholder]

**Application**

[Application to facts, with element-by-element analysis where applicable]

**Conclusion on Issue 2**

[Tentative conclusion with confidence level]

---

*[Add additional Issue sections as needed.]*

---

### Adverse Authority and Counterarguments

- [Adverse authority or counterargument 1, with citation / source, and brief note on why it does or does not control]
- [CONFIRM: attorney should independently search for adverse authority not identified in this session before relying on this memo]

---

## Conclusion

[Question 1]: [Summary answer with confidence level and key dependencies.]

[Question 2, if applicable]: [Summary answer with confidence level and key dependencies.]

**Key variables:** [List the two to four facts or legal determinations that most affect the outcome. Attorney should focus verification effort here first.]

---

## Authorities Cited

| # | Authority | Type | Citation | Proposition asserted | Source | Verified by attorney |
|---|-----------|------|----------|----------------------|--------|----------------------|
| 1 | [Case / Statute / Regulation / Secondary Source name] | [Case / Statute / Reg / Secondary] | [Full citation] | [Specific holding or rule this authority supports] | [User-provided / Researched — identify source] | - [ ] |
| 2 | | | | | | - [ ] |
| 3 | | | | | | - [ ] |
| 4 | | | | | | - [ ] |
| 5 | | | | | | - [ ] |

**Authority source key:**
- **User-provided** — document or citation supplied directly by the user or client.
- **Researched** — retrieved from a named research source during this session; identify the source.
- **Connector-verified** — citation resolved through a documented connector (e.g., CourtListener — see `connectors/courtlistener.md`); record the verified opinion URL in the Source column. The citation form is confirmed; whether the cited authority supports the asserted proposition remains an attorney verification item.
- **[VERIFY-CITE]** / **[CONFIRM]** — authority has not been verified; attorney must locate, review, and confirm before relying on this memo.

---

## Open Items / Attorney Verification

- [ ] Confirm jurisdiction and governing law are correctly identified for this matter.
- [ ] Confirm the relevant date (triggering event, limitations period start, statute effective date) is correct.
- [ ] Verify all case citations exist and that the holdings attributed to them are accurate.
- [ ] Verify all statutes and regulations are in their current version and say what this memo claims.
- [ ] Independently search for adverse authority not identified in this session.
- [ ] Resolve all `[CONFIRM: ...]` placeholders throughout the memo.
- [ ] Confirm all facts stated in the Facts section are accurate and complete.
- [ ] Resolve factual gap: [describe gap from Facts section, or delete if none]
- [ ] Confirm assumption: [describe assumption from Assumptions section, or delete if none]
- [ ] Assess conflict-of-laws or choice-of-law issues, if implicated.
- [ ] Assess whether any procedural deadline is affected by this legal analysis.
- [ ] Confirm that the privilege and confidentiality designations are appropriate for how this memo will be circulated.
- [ ] Attorney to confirm overall conclusions and confidence levels before client communication or reliance.

---

=== END AGENTCOUNSEL MODULE: legal-research-memo-template ===

=== BEGIN AGENTCOUNSEL MODULE: citation-type-taxonomy ===
Kind: reference
Canonical path: skills/legal-research/references/citation-type-taxonomy.md
Load when: the memo uses multiple authority types or citation forms

> Shared reference material supporting the AgentCounsel legal-research and legal-methodology skills, used to help produce draft legal work product for attorney review — not legal advice.

# Citation Type Taxonomy

This reference gives a reviewer a vocabulary for classifying what *kind* of
citation a passage in a draft actually is, before deciding what verifying it
would require. Different citation forms carry different verification needs
and different hallucination risks — an `id.` is only meaningful in relation
to the citation before it, while a full case citation can be checked in
isolation. This taxonomy does not verify anything itself; it tells a skill
which question to ask for each citation form it finds.

**Attribution.** The set of citation-type categories below is adapted from
the classification scheme used by eyecite, an open-source citation-extraction
library maintained by the Free Law Project, licensed BSD-2-Clause. The
descriptions of each type, the verification needs, and the hallucination-risk
notes are written independently for AgentCounsel; they are not eyecite's
code, docstrings, or documentation text.

---

## How Skills Use This File

`skills/legal-methodology/citation-integrity-check/SKILL.md` and
`skills/legal-methodology/source-validation/SKILL.md` classify each citation
in a draft before assessing its source status or integrity. This taxonomy is
the vocabulary for that classification step — it identifies *what form* a
citation takes; the two skills above still own the questions of *whether it
is real*, *whether it is complete*, and *whether it supports the proposition
asserted*. Legal-research skills that assemble authority-heavy drafts (for
example `skills/legal-research/legal-research-memo/SKILL.md` and
`skills/legal-research/authority-synthesis/SKILL.md`) can also use this
taxonomy while drafting, to keep short-form citations correctly anchored to
the full citation they depend on.

---

## Citation Types

### Full case citation

**What it looks like:** A complete citation to a court decision — case name,
reporter volume, reporter abbreviation, starting page (and pin cite if
applicable), court, and year, e.g. a pattern like `Party A v. Party B, 123
F.3d 456, 460 (9th Cir. 1999)`.

**What a verifier needs:** The citation is self-contained — no earlier
citation in the document is required to resolve it. Verification means
confirming the reporter abbreviation exists (see
`connectors/reporters-and-courts.md`), that the court abbreviation matches a
real court, and that the case can be located under that name and citation.

**Hallucination risk:** This is the highest-risk form, because a fabricated
full citation can look completely well-formed — a plausible party-name
pattern, a real reporter abbreviation, and a page number within a plausible
range, none of which make the case real. A citation "looking right" is not
evidence it exists.

### Short-form case citation

**What it looks like:** An abbreviated citation used after a case has already
been cited in full, typically the case name plus a shortened reference to
the reporter and page, e.g. `Party A, 123 F.3d at 462`.

**What a verifier needs:** The preceding full citation for the same case
somewhere earlier in the document. A short form cannot be verified in
isolation — it is only correct if it actually matches an antecedent full
citation for the same case in the same document.

**Hallucination risk:** A short form can silently drift from its antecedent
— citing a different page, or a short form with no matching full citation
anywhere in the document at all (an orphaned short form), which is itself a
sign the full citation was never actually verified or was dropped in
editing.

### Supra citation

**What it looks like:** A citation using "supra" to refer back to a source
cited earlier by name, without repeating the reporter citation, e.g. `Party
A, supra, at 462` or `Smith, Legal Treatise, supra note 4, at 12`.

**What a verifier needs:** The specific earlier citation or footnote the
supra reference points to. A supra reference is meaningless without
resolving exactly which prior citation it targets — this often requires
matching a footnote number or a uniquely identifying short form.

**Hallucination risk:** A supra reference to a footnote or citation number
that does not exist in the document, or that points to the wrong source
once resolved, silently misattributes a proposition to a citation that never
actually supports it.

### Id. citation

**What it looks like:** A citation of `Id.` or `Id. at [page]`, referring to
the same source as the immediately preceding citation.

**What a verifier needs:** The **immediately preceding** citation in the
document — not the most recent citation of that source anywhere in the
draft, but the one immediately before it. An `Id.` is only correct if nothing
else was cited in between.

**Hallucination risk:** `Id.` is the form most vulnerable to silent drift
during editing — inserting, deleting, or reordering a citation elsewhere in
the document can leave a surviving `Id.` pointing to the wrong source without
any visible change to the `Id.` text itself. Every `Id.` must be re-checked
against its immediately preceding citation whenever the surrounding text is
edited.

### Statutory citation

**What it looks like:** A citation to a statute or code section, e.g. a
pattern like `15 U.S.C. § 78j(b)` or a state-code citation, typically title
or chapter, section symbol, and section number.

**What a verifier needs:** Confirmation of the code, title/chapter, and
section number, and — critically — the **version or effective date**, since
statutory text changes over time and a citation with no version marker is
ambiguous about which version of the law is being cited.

**Hallucination risk:** A plausible-looking but non-existent section number,
or a real section number whose current text has been amended since the date
implicitly assumed by the draft. A statutory citation without an "as of"
date is not verifiable as currently stated law without independent research.

### Regulation citation

**What it looks like:** A citation to a codified administrative regulation,
e.g. a pattern like `17 C.F.R. § 240.10b-5`, or a citation to an agency's
Federal Register document by volume and page.

**What a verifier needs:** The same version-date discipline as a statutory
citation, plus confirmation of which agency issued it — regulations are
amended, and a codified regulation citation with no effective-date anchor
cannot be assumed current.

**Hallucination risk:** A fabricated part or section number under a real
title, or a citation to a rule that has since been superseded, withdrawn, or
stayed without the draft acknowledging that possibility.

### Law-journal citation

**What it looks like:** A citation to a law review or journal article, e.g.
a pattern like `Author Name, Article Title, 100 Some L. Rev. 1 (2020)`.

**What a verifier needs:** The author, title, journal volume/abbreviation,
starting page, and year — all self-contained, similar to a full case
citation, but for secondary rather than primary authority. Confirming
existence does not confirm the article actually supports the proposition
cited to it.

**Hallucination risk:** A fabricated author, title, or journal issue that
reads as entirely plausible — law review citations are especially easy to
invent convincingly because there is no small, checkable universe of
reporters or courts to match against, unlike case citations.

### Unknown / partial citation

**What it looks like:** A citation fragment that does not resolve cleanly
into any of the categories above — a case name with no reporter cite, a
section reference with no code identified, a citation copied with a typo, or
a citation abbreviated so heavily that its type cannot be determined from the
text alone.

**What a verifier needs:** More context before any verification step is even
possible. Do not guess at the missing elements or "complete" the citation
from model background knowledge — treat it as an incomplete citation and
flag it for the source to be supplied or the citation corrected.

**Hallucination risk:** The temptation to silently "fill in" a plausible
missing reporter, section number, or year is the single highest-risk moment
in citation handling. An unknown/partial citation must never be completed
from memory; completing it is how a hallucinated citation gets introduced.

---

## Verification Workflow Per Type

| Citation type | Resolves independently? | What must be checked before verification is even possible | Primary hallucination risk |
|---|---|---|---|
| Full case citation | Yes | Reporter and court abbreviation exist (`connectors/reporters-and-courts.md`); case locatable under that name and cite | Entirely fabricated but well-formed citation |
| Short-form case citation | No — needs its antecedent full citation | Locate and match the earlier full citation for the same case | Drift from, or no match to, the antecedent |
| Supra citation | No — needs the specific earlier reference | Locate and match the exact footnote/citation the supra points to | Wrong or missing target reference |
| Id. citation | No — needs the immediately preceding citation | Confirm nothing else was cited in between | Silent drift after document edits |
| Statutory citation | Yes, with a version caveat | Code, section number, and version/effective date | Fabricated section number; stale version |
| Regulation citation | Yes, with a version caveat | Title, part/section, issuing agency, effective date | Superseded, withdrawn, or stayed rule cited as current |
| Law-journal citation | Yes | Author, title, journal/volume, page, year | Fabricated author, title, or issue |
| Unknown / partial | No | Additional source material or correction from the user | Silently "completing" the citation from memory |

---

## Reviewer Notes

- This taxonomy classifies citation *form*. It does not substitute for the
source-status classification in
`skills/legal-methodology/source-validation/SKILL.md` or the integrity
checks in `skills/legal-methodology/citation-integrity-check/SKILL.md` —
use this file to identify the type first, then apply those skills'
workflows.
- For existence checks on case-law reporter abbreviations and court names,
see `connectors/reporters-and-courts.md`. Existence of a reporter or court
abbreviation is not existence of the case itself — see that connector's
scope and limits.
- An unresolved short-form, supra, or `Id.` citation (no matching antecedent
found in the document) is itself a defect to flag — do not assume the
antecedent existed and was simply omitted from view.

=== END AGENTCOUNSEL MODULE: citation-type-taxonomy ===

=== BEGIN AGENTCOUNSEL MODULE: courtlistener-connector-guide ===
Kind: connector
Canonical path: connectors/courtlistener.md
Load when: US federal case existence or citation verification is available through CourtListener

# CourtListener

> Reference material supporting the AgentCounsel skill library, used to help produce draft legal work product for attorney review — not legal advice.

This connector points at [CourtListener](https://www.courtlistener.com), the [Free Law Project](https://free.law)'s open database of US case law. It is the first concrete connector under `connectors/`; the broader framing is in [`README.md`](README.md).

The Free Law Project explicitly frames the Citation Lookup endpoint documented here as a guardrail "to help prevent hallucinated citations" (source: [Citation Lookup and Verification API](https://www.courtlistener.com/help/api/rest/citation-lookup/)). That is the use case this connector enables for an AgentCounsel skill — closing the gap between "I left a `[VERIFY-CITE: ...]` placeholder" and "I confirmed the case exists with that citation in a reliable corpus."

## 1. Source

- **Publisher:** Free Law Project (a US-based nonprofit). CourtListener is a research-and-disclosure project, not a commercial legal research vendor.
- **Cost:** Free for read access. No API key required to begin; an account is recommended for monitoring and higher tiers.
- **API version:** Current is **v4** (v4.4 at the time this doc was written). The base URL is `https://www.courtlistener.com/api/rest/v4/`. Older versions exist but should not be relied on for new work; consult the [V4 Migration Guide](https://www.courtlistener.com/help/api/rest/v4/migration-guide/) before using anything older. `[CONFIRM: API version currency before integrating in a long-lived tool]`.
- **Rate limits** (per the Free Law Project's published limits at the time this doc was written): up to **5,000 requests per hour** for unauthenticated users; **5,000 requests per day** for free accounts; the API documentation also notes a **60-requests-per-minute** throttle. Limits are noted as in flux pending a new membership model. `[CONFIRM: current rate limits before relying on a sustained workflow]`. Source: [REST API overview](https://www.courtlistener.com/help/api/rest/) and the Free Law Project's [discussion on limits](https://github.com/freelawproject/courtlistener/discussions/6895).
- **License of returned content:** CourtListener opinions are US public-domain court records; the Free Law Project's own metadata and database are licensed by them under permissive terms. `[CONFIRM: applicable Free Law Project terms before redistribution]`.

This connector documents the **free, mostly no-key surface**. CourtListener also exposes PACER-derived data and paid features; both are out of scope here.

## 2. In scope — what CourtListener can verify

CourtListener's corpus and the surface this connector documents support verifying:

- **US Supreme Court opinions** — full historical corpus.
- **US Courts of Appeals** — federal appellate decisions across all circuits.
- **US District Courts** — selective coverage of district-court opinions.
- **State courts of last resort** — varying coverage by state; many state supreme courts are present, with coverage lag and gaps that vary by jurisdiction. Treat state coverage as partial unless a specific state has been confirmed.
- **Bankruptcy, tax, and specialty federal courts** — partial coverage.
- **Per-opinion basics** — case name, citation(s), court, date, judges, and full opinion text where the opinion is published.
- **Citator-style relationships** — what other opinions cite a given opinion (the "OpinionsCited" graph).
- **Reverse citation lookup** — given a citation string, locate the matching opinion(s). The Citation Lookup endpoint parses citations using [Eyecite](https://github.com/freelawproject/eyecite), which the project says was developed against more than 50 million citations going back more than two centuries.

What "verify" means here: confirm that the cited case **exists**, that the **citation form matches** what CourtListener has on record, and that the **public URL** for the opinion can be recorded for attorney review. **Verifying that the opinion stands for the proposition for which it is cited is not the connector's job — that is attorney work.**

## 3. Out of scope — what CourtListener does not cover

A skill must keep its existing placeholder discipline for anything in this list:

- **Many state intermediate appellate and trial court opinions** — coverage is partial and varies by state.
- **Unpublished or sealed proceedings** — not in the public corpus.
- **Very recent decisions** — coverage lag varies; new opinions appear on a delay.
- **Secondary sources** — treatises, restatements, law review articles, practice guides.
- **Statutes, regulations, court rules, and agency guidance** — CourtListener's primary focus is case law. Statutory verification belongs to a different connector.
- **Foreign and international tribunals** — out of scope. Use a jurisdiction-specific connector when one is added.
- **Pinpoint accuracy of internal page numbers in older volumes** — verify pinpoint cites against the official reporter.

If a skill needs to verify any of the above, treat the connector as unavailable for that placeholder and retain the existing `[VERIFY-CITE: ...]` / `[CONFIRM: ...]` flag.

## 4. Surface — where to hit

### 4a. Public web search (no API)

The simplest surface, available in any browser or HTTP-capable tool:

- **Search URL pattern:** `https://www.courtlistener.com/?q=<URL-encoded-query>`
- **Opinion page URL pattern:** `https://www.courtlistener.com/opinion/<opinion-id>/<slug>/`

Use this surface for: quick visual confirmation of a case, sharing a stable URL with the attorney, or environments without HTTP tooling.

### 4b. Citation Lookup API — the primary verification surface

The most reliable way to resolve a citation string to a specific opinion:

- **Endpoint:** `https://www.courtlistener.com/api/rest/v4/citation-lookup/`
- **HTTP method:** `POST` (not GET).
- **Input modes (per the [Citation Lookup docs](https://www.courtlistener.com/help/api/rest/citation-lookup/)):** the API can look up either a single individual citation or parse and look up every citation in a block of text.
- **Hard limits:** at most **250 citations per request** (excess parsed but returned with a `status` value indicating "Too many citations requested"); at most **64,000 characters** of text per request.
- **Response shape (per the official docs):** each match returns a `citation` (the looked-up form), `normalized_citations` (canonical form(s); multiple entries when ambiguous), `start_index` and `end_index` (positions of the citation in the input text), and a `status` field (`200` indicates a found and valid match).
- **Authentication:** uses the same auth and serialization as the rest of the CourtListener API; works unauthenticated subject to the rate limits in Section 1.

This is the endpoint a skill should hit when resolving a `[VERIFY-CITE: ...]` placeholder. The Eyecite-backed parsing handles ambiguous and non-canonical citation forms that a plain string match would miss.

### 4c. Search API

For richer queries (court filter, date range, case-name search):

- **Endpoint:** `https://www.courtlistener.com/api/rest/v4/search/`
- **HTTP method:** `GET` with query parameters, e.g., `?q=<query>`.
- **Supported filters** (consult the live [Search API docs](https://www.courtlistener.com/help/api/rest/search/) for the current list): query text, document type, court filter, date filed range. `[CONFIRM: exact filter parameter names against the live docs — they change with API minor versions]`. The Free Law Project recommends sending an `OPTIONS` request to any endpoint to discover the current filter set.

Use this for: locating a case by partial information when only a name or topic is known, narrowing within a specific court or date window.

### 4d. Opinion fetch

For pulling the full text of a verified opinion:

- **Endpoint pattern:** `https://www.courtlistener.com/api/rest/v4/opinions/<opinion-id>/`
- **HTTP method:** `GET`.
- **Use for:** confirming that a paraphrase or quotation matches the source text. Confirming the quotation exists in the opinion is a structural check; the legal weight of the quoted passage remains an attorney determination.

### 4e. MCP tool surface

When the user's environment includes an MCP server that wraps CourtListener (community wrappers exist; check the user's installed MCP servers), prefer the MCP tool surface over raw URL fetches — it handles authentication, retries, and rate-limit backoff. The connector contract is the same: query, record the verified URL, fall back to placeholder if unavailable. **This connector does not configure or install any MCP server** — see `connectors/README.md`.

## 5. Calling pattern from a skill

When a skill's Workflow reaches a citation it would otherwise mark `[VERIFY-CITE: ...]` or `[CONFIRM: ...]`, and the citation is in scope under Section 2:

1. **Look up the citation.** For a citation string in a draft, `POST` to the Citation Lookup endpoint (4b). For a case-name or topic-based lookup, use the Search endpoint (4c).
2. **Branch on the result:**
- **Exact match (single result):** Record the case in the Authorities Cited table (or equivalent) with the CourtListener opinion URL appended, e.g., `Erie R.R. Co. v. Tompkins, 304 U.S. 64 (1938) — verified via CourtListener: https://www.courtlistener.com/opinion/<id>/erie-railroad-co-v-tompkins/`. Replace the `[VERIFY-CITE: ...]` placeholder with `[ATTORNEY TO CONFIRM: proposition supported by the cited case]` — the citation now exists; whether it stands for the asserted proposition is still attorney work.
- **Multiple matches:** Do not pick one. Record all matches, retain the `[VERIFY-CITE: multiple matches in CourtListener — attorney to select]` flag, and escalate.
- **No match:** Retain a `[VERIFY-CITE: not found in CourtListener — attorney to verify in another source]` flag. Do not delete the asserted citation; the attorney needs to see what was claimed.
- **Out of scope (per Section 3):** Keep the original placeholder. Do not query CourtListener for state intermediate appellate cases or non-case-law authority and treat absence-of-result as a verification.
3. **If the connector is unavailable** (no MCP tool, no HTTP capability, rate-limit exceeded): retain the original placeholder unchanged. Add `[VERIFY-CITE: not verified — no CourtListener connector available in this session]` so the attorney sees the gap.

The Output Format of any skill that consumes this connector should record the verification source per authority — see the per-skill changes in `skills/legal-research/legal-research-memo/SKILL.md`, `skills/litigation/brief-section-drafter/SKILL.md`, and `skills/litigation/claim-chart/SKILL.md`.

## 6. Fallback behavior

| Outcome | Action |
|---|---|
| Single match | Record verified URL; mark `[ATTORNEY TO CONFIRM: proposition supported]` |
| Multiple matches | List all; mark `[VERIFY-CITE: multiple matches — attorney to select]` |
| No match | Mark `[VERIFY-CITE: not found in CourtListener — attorney to verify elsewhere]` |
| Out of scope per Section 3 | Retain the original placeholder; do not query |
| Connector unavailable | Retain the original placeholder; add `not verified — no CourtListener connector` note |
| Rate-limit reached | Treat as unavailable for the remainder of the session; flag in Open Items |

## 7. Limits and known failure modes

- **Coverage lag.** New opinions appear with a delay. A citation that does not appear today may appear next week. Do not infer that absence in CourtListener means a case was not decided.
- **State coverage is partial.** A negative result for a state court should be read as "not in CourtListener," not as "does not exist."
- **Citation form drift.** Older citations, parallel citations, and unconventional formats can return false negatives. The Citation Lookup endpoint's Eyecite parser handles many such forms; for stubborn cases, try both the volume-page form (`304 U.S. 64`) and the case name (`Erie v. Tompkins`).
- **Pinpoint citations.** Confirming the case exists is not confirming a pinpoint page or paragraph. Pinpoints remain attorney-verification items.
- **Cited-but-not-quoted propositions.** A case being real does not mean the proposition asserted in the draft is supported by the case. That determination remains with the attorney.
- **Versioning.** Opinions can be revised; en banc rehearings and amended opinions can change pin cites. Record the date of the verification and re-verify before reliance on a sensitive citation.
- **API surface drift.** The Free Law Project is actively developing v4; minor versions add and rename filters. Treat the URLs and parameter names above as the shape *at the time this doc was written* and re-check the live docs (`/help/api/rest/`) before automating against them.

## 8. What this connector does not do

- Does not verify statutes, regulations, court rules, or secondary sources.
- Does not validate that a case is still good law (Shepardizing / KeyCiting is a separate operation that this free surface does not perform).
- Does not characterize a holding, the level of generality of a holding, or the precedential weight of a case in a particular court.
- Does not assess whether a case is binding, persuasive, or distinguishable in the operative forum.

All of those remain attorney-verification items. The connector closes the "does this case exist with this citation" question. Nothing more.

=== END AGENTCOUNSEL MODULE: courtlistener-connector-guide ===

=== BEGIN AGENTCOUNSEL MODULE: negative-treatment-workflow ===
Kind: workflow-module
Canonical path: skills/legal-research/negative-treatment-check/SKILL.md
Load when: deep-review mode or when a material case controls the analysis

---
name: Negative Treatment Check
description: "Use when you need to check whether the authorities cited in a draft or research memo are still good law — organizing a citator-style verification plan for each authority, recording provided citator-report signals as attributed claims, and classifying each authority's verification status, without ever asserting from model memory that a case is good law, overruled, or distinguished."
practice_area: legal-research
task_type: analysis
jurisdictions: []
risk_level: high
requires_attorney_review: true
inputs:
- "The draft, memo, or citation list whose authorities need treatment verification"
- "For each authority: the full citation as it appears in the source"
- "Optional: any citator reports or validation printouts the user has already obtained"
- "Optional: access to a citation-verification connector (e.g., connectors/courtlistener.md) for federal case law"
outputs:
- "Authority-by-authority treatment-verification table"
- "A verification plan for each authority, using available connectors and user-supplied citator reports"
- "Verification-status classification and attorney sign-off items"
related_skills:
- skills/legal-research/authority-synthesis/SKILL.md
- skills/legal-research/legal-research-memo/SKILL.md
- skills/legal-methodology/citation-integrity-check/SKILL.md
tags:
- legal-research
- citator
- negative-treatment
- good-law
- citation-verification
- source-validation
---

# Negative Treatment Check

## Purpose

Organize a disciplined "is this still good law" check for every authority cited in a draft or research memo. For each authority, the skill records the citation as provided, structures a verification plan using available connectors and any user-supplied citator reports, and classifies the authority's verification status. It never asserts from model memory that a case is good law, has been overruled, superseded, limited, or distinguished — model-memory treatment claims are exactly the hallucination this library exists to prevent. Every treatment conclusion is either drawn from a provided, attributed source or left as an explicit `[Verify current law]` item for attorney confirmation. This is draft legal work product for attorney review, not legal advice.

## Use When

- A research memo or brief is nearly final and every cited authority needs a good-law check before it is filed or relied upon.
- A user asks "is this case still good law?" or "have any of these authorities been overruled?" and wants the verification organized rather than guessed.
- A citation list has been assembled and the team needs a treatment-verification plan before an attorney signs off.
- Citator reports have been pulled from a vendor service and the signals need to be organized into a per-authority status table.
- A cited authority is old, from a fast-moving area, or central to the argument, and its current status must be confirmed.

## Required Inputs

- **The draft, memo, or citation list** whose authorities need checking, with each authority's full citation as it appears in the source. If only a description is provided, request the actual citations — never reconstruct a citation.
- Optional: **any citator reports or validation printouts** the user has already obtained from a citation service. These are recorded as attributed claims from a provided source.
- Optional: **access to a citation-verification connector** (for example, `connectors/courtlistener.md` for federal case law) — used to confirm that an authority exists and to retrieve its text and later history where the connector supports it.

If no citations are provided, stop and request them. Do not generate or assume citations to check.

## Do Not Use When

- The task is to synthesize authorities into a rule or to build the substantive analysis — use `authority-synthesis`; run this check after synthesis, before finalizing.
- The task is to verify that citations are formatted correctly and that quotations match their sources — use `citation-integrity-check`; this skill checks treatment (still-good-law status), not citation form.
- The task is to draft the research memo itself — use `legal-research-memo`.
- The user wants a definitive statement that a case is or is not good law without any provided citator source — this skill cannot supply that; it organizes verification and flags the gap.
- No authorities are cited — there is nothing to check.

## Legal Safety Rules

- **Source and citation discipline.** Follow `core/source-and-citation-discipline.md`. Never invent legal authority, citations, quotations, statutes, cases, regulations, filing deadlines, or procedural rules. Label what is a provided source, a user-provided fact, an assumption, a legal inference, or an item requiring attorney verification, and use a citation placeholder such as `[Attorney to insert authority]` when no source is available.
- Produce draft legal work product for attorney review. This is not legal advice.
- **Never assert treatment from model memory.** Do not state — from your own knowledge — that any authority is good law, has been overruled, reversed, superseded, abrogated, limited, criticized, or distinguished. Treatment claims not drawn from a provided, attributed source are the exact hallucination risk this skill exists to contain. State this limitation plainly in the output.
- **Record vendor citator signals as attributed claims, never as confirmed facts.** When a user supplies a citator report, record what the report states (its flags, its listed citing references, its date) attributed to that report — never independently "confirm" or upgrade it. Note that citator services can lag and can miss treatment; the attorney verifies.
- **Connector results verify existence and history, not legal status.** A connector can confirm an opinion exists, retrieve its text, and surface later citing documents where supported — it does not adjudicate whether the authority remains good law. Frame connector findings accordingly.
- Classify every authority's status explicitly, and default to `[Verify current law]` whenever the status rests on anything other than a provided, attributed citator source.
- Never compute or assert a date. Record the dates a provided report or connector states, attributed to the source.
- Distinguish throughout: what a provided report or connector states (attributed), what the user stated, what is assumed, and what the attorney must verify.
- Preserve confidentiality: keep the matter's authorities and any work-product framing out of reusable templates.

## Workflow

1. **Confirm inputs.** Verify the citation list (or the draft's authorities) is provided with full citations. Note which citator reports the user supplied and whether a verification connector is available. Request anything missing.
2. **Extract the authority list.** List every authority cited in the source — cases, statutes, regulations, rules, and secondary sources — with its full citation exactly as it appears and the proposition it is cited for. Flag any citation that is incomplete or malformed as `[CONFIRM: citation]` (and route form issues to `citation-integrity-check`).
3. **State the limitation up front.** Open the working notes with the explicit statement that no treatment status in this output is asserted from model memory, and that every "current" status rests on a provided citator source or an attorney's own verification.
4. **Plan verification for each authority.** For each authority, record the verification path: which supplied citator report covers it; whether a connector (e.g., CourtListener for federal case law) can confirm existence and retrieve later history; and what the attorney will still need to run in a full citator service. Note where no verification source is available at all.
5. **Record provided citator signals.** Where the user supplied a citator report, record for each authority: the report's treatment flag or signal (as the report labels it), the citing references it lists, and the report's date — each attributed to the report. Do not translate a vendor's caution flag into a legal conclusion.
6. **Record connector findings.** Where a connector was used, record that the authority exists (or that it could not be located — a prominent flag), its retrieved metadata, and any later citing documents the connector surfaced, each attributed to the connector. Do not infer good-law status from the presence or absence of citing documents.
7. **Note high-risk authorities.** Flag authorities that warrant heightened verification: old cases, cases in fast-moving or recently-legislated areas, cases central to the argument, and any authority relied on for a proposition broader than its holding (route the proposition-scope question to `authority-synthesis`).
8. **Classify each authority's verification status.** Assign one of: **Verified current (attorney-confirmed)** — only where an attorney has confirmed; **Citator report provided** — status per an attributed report, attorney to confirm; **Connector-existence-only** — existence confirmed, treatment not; **Unverified** — no treatment source, treat as `[Verify current law]`. Default to Unverified when in doubt.
9. **Assemble the treatment-verification table** and list, prominently, every authority that is Unverified or that a provided report flags for attention.
10. **List attorney verification items and assemble the output.** Consolidate placeholders, assemble the output in the format below, label it a draft for attorney review, and attach the checklist.

## Output Format

Deliver the following, in order, labeled `DRAFT — For Attorney Review — Treatment Not Confirmed From Model Knowledge`:

1. **Summary** — one paragraph: how many authorities were checked, how many remain Unverified, and the explicit statement that no treatment status is asserted from model memory.
2. **Verification Limitation Notice** — the plain statement from Step 3.
3. **Treatment-Verification Table** — columns: Authority (full citation) | Cited For | Verification Source (report / connector / none) | Reported Signal (attributed) | Status Classification | Flags.
4. **High-Risk Authorities** — the authorities from Step 7 needing heightened verification, with the reason.
5. **Unverified Authorities** — every authority with no treatment source, each `[Verify current law]`.
6. **Verification Plan** — for each authority, what the attorney must still run to confirm status.
7. **Attorney Verification Items** — every placeholder consolidated.
8. **Assumptions** — every assumption made, listed explicitly.

Use `[Verify current law]` on every status not resting on an attorney's own confirmation. Do not fill gaps with invented content.

## Attorney Verification Checklist

- [ ] Every authority in the source has been captured, with a complete, correct citation.
- [ ] No treatment status in this draft was asserted from model knowledge; each status rests on a provided citator source or attorney verification.
- [ ] Every authority has been run through a full citator service by the attorney or a qualified researcher.
- [ ] Any authority a provided report flagged for negative or cautionary treatment has been read and assessed in context.
- [ ] Authorities central to the argument have been independently confirmed as current good law.
- [ ] For each authority, the proposition it supports has been checked against its actual holding, not just its good-law status.
- [ ] Connector-existence-only authorities have had their treatment confirmed through a citator, not inferred from citing-document counts.
- [ ] Old authorities and authorities in fast-moving or recently-amended areas have received heightened verification.
- [ ] Statutes and regulations have been checked for currency against the official code, not only through case citators. `[Verify current law]`
- [ ] Every Unverified authority has been resolved or removed before the draft is relied upon.
- [ ] No date in this draft was computed; all dates are attributed to a provided report or connector.
- [ ] The citation form of every authority has been separately verified (via `citation-integrity-check`).
- [ ] All `[Verify current law]` and `[CONFIRM: ...]` placeholders have been resolved before the memo or brief is filed or relied upon.

=== END AGENTCOUNSEL MODULE: negative-treatment-workflow ===

=== BEGIN AGENTCOUNSEL MODULE: source-validation-workflow ===
Kind: quality-check
Canonical path: skills/legal-methodology/source-validation/SKILL.md
Load when: standard or deep-review mode

---
name: Source Validation
description: "Use when classifying whether cited sources, authorities, quotations, dates, and factual claims in draft legal work product are supported by materials available in the session, applying the source hierarchy in `core/source-and-citation-discipline.md`."
practice_area: legal-methodology
task_type: verification
jurisdictions: []
risk_level: medium
requires_attorney_review: true
inputs:
- "The draft legal work product to check"
- "The cited sources, authorities, and underlying record"
- "The claims, quotations, and dates to validate"
outputs:
- "Claim table classifying support status for factual and legal claims"
- "Source reference table"
- "Unsupported claims and contradictions list"
related_skills:
- skills/legal-methodology/red-team-verifier/SKILL.md
- skills/legal-methodology/citation-integrity-check/SKILL.md
- skills/legal-methodology/assumption-audit/SKILL.md
- skills/legal-research/legal-research-memo/SKILL.md
tags:
- legal-methodology
- source-validation
- citation-check
- verification
---

# Source Validation

## Purpose

This source-validation check now uses a claim-by-claim support taxonomy. For each material factual or legal claim, classify support as **source-supported**, **source-mentioned but insufficient**, **unsupported**, **contradicted by source**, **legal authority required**, or **attorney judgment required**. These classifications organize attorney review; they do not independently verify current law or certify that a legal conclusion is correct.

Apply a systematic method for checking that every cited source, authority, quotation, date, and factual claim in a draft legal work product actually exists and says what is claimed. The skill enumerates every citation and claim in the draft, classifies each by source tier using the hierarchy in `core/source-and-citation-discipline.md`, and marks each as verified, unverified, or unverifiable. It produces a source-by-source validation table and a list of unresolved items that must be checked before the draft is relied upon.

## Use When

- A draft legal work product contains citations, quotations, authority references, or factual claims that need to be checked before the draft is finalized or passed to attorney review.
- A user asks to "verify the sources," "check the citations," "confirm these quotes are real," or "validate the authority in this draft."
- The Red-Team Verifier (or any review process) has flagged specific citations or claims as unverified and a dedicated source check is needed.
- A legal research memo, brief section, or analysis contains cases, statutes, regulations, or secondary sources that were gathered during research and need to be confirmed before attorney reliance.
- A draft quotes from a source document (contract, filing, policy, statute) and the quotations need to be checked against the provided text.
- The user wants a complete audit trail showing which sources in a draft are confirmed and which require additional verification.

**Choosing among the verification skills:**

- Use this skill for claim-by-claim classification of whether each claim in a draft is supported by the available sources.
- Use `red-team-verifier` instead for the full adversarial pass over an entire deliverable — the superset that also covers reasoning, framing, jurisdiction, and completeness.
- Use `hallucination-red-team` instead for a targeted hunt for fabricated or unsupported content in a draft.
- Use `citation-integrity-check` instead when the review is limited to citations and quotations.

## Required Inputs

- The complete draft to be validated (uploaded or pasted). Do not validate from a summary or paraphrase.
- The source documents or authorities that are supposed to support the claims in the draft — for example, the contract the draft reviews, the cases the memo cites, or the regulatory text the analysis references. These are the materials against which quotations and propositions are checked.
- Optional: the skill or workflow that produced the draft — this informs which sourcing standards apply.
- Optional: a list of specific citations or claims flagged for priority review.

## Do Not Use When

- The task is to conduct new legal research and find sources for unsupported propositions — use `legal-research-memo` or a research skill; this skill validates existing sources, it does not find new ones.
- The task is to conduct a broader structural review of draft quality — use the Red-Team Verifier for that purpose; use this skill for focused source and citation checking.
- The draft contains no citations, quotations, or factual claims traceable to external sources — there is nothing to validate.
- The user needs an attorney to assess whether an authority supports a legal argument — that is a substantive judgment beyond the scope of this skill; flag those items as attorney verification items.

## Legal Safety Rules

- Produce draft legal work product for attorney review. This is not legal advice.
- Model background knowledge is never a verification source. A citation that appears consistent with what the model knows about a case or statute is still classified as unverified unless it is present in a user-provided document or independently retrieved and confirmed in this session.
- Do not assert that a citation is valid or that a quotation is accurate based on model background knowledge, training data, or general familiarity with legal sources.
- Do not invent, reconstruct, or complete a partial citation. If a citation is incomplete, mark it as incomplete and list it as an unresolved item.
- Do not assert that a case, statute, or regulation does not exist — only that it could not be verified within this session. Absence of verification is not confirmation of fabrication; that determination requires independent research.
- Do not assess whether an authority is controlling, persuasive, current, or correctly applied — those are attorney functions. This skill checks existence and textual accuracy, not legal significance.
- Do not place client-sensitive facts from the draft into any reusable output beyond this validation report.
- Label the validation report as draft legal work product for attorney review.

## Workflow

1. **Confirm inputs.** Verify that you have the complete draft text. Note whether source documents have been provided and, if not, record the resulting limitation: without source documents, only citation-tier classification is possible — quotation checking and claim-source tracing require the underlying documents.

2. **Enumerate every citation and claim.** Use `skills/legal-research/references/citation-type-taxonomy.md` to classify each citation's form (full case citation, short-form, supra, `Id.`, statutory, regulation, law-journal, or unknown/partial) so that short forms, supras, and `Id.` citations are correctly matched to their antecedent before verification. Also assign each citation a verification tier using `skills/legal-methodology/references/citation-confidence-tiers.md` alongside the source-tier classification in step 3 below, so the report distinguishes citation form from verification status. Read through the draft systematically and build a complete inventory of:
- Case citations: case name, citation, reporter, year, court.
- Statutory and regulatory citations: instrument name, section, version or effective date.
- Secondary source citations: treatise, article, report, guidance document.
- Quotations: every passage presented as a direct quotation from any source.
- Factual claims: every claim of fact that depends on a source document, client representation, or external authority — as distinct from analytical inference.
- Dates: every specific date asserted in the draft, including effective dates, filing dates, events, and deadlines.

Assign each item a unique number for tracking. If the draft is long, work section by section and flag the section with each entry.

For every material factual or legal claim, also assign one source-support status: **Source-supported**, **Source-mentioned but insufficient**, **Unsupported**, **Contradicted by source**, **Legal authority required**, or **Attorney judgment required**. Use "source-supported" only when an available source directly supports the claim as stated, subject to attorney review.

3. **Classify each item by source tier.** Apply the hierarchy from `core/source-and-citation-discipline.md`:
- **Tier 1 — User-provided document.** The claim or quotation is traceable to a document the user supplied in this session. This is the highest-trust tier; verify by checking the text of the provided document.
- **Tier 2 — Independently researched and verified.** The claim was located through a research step and confirmed to exist and to say what is claimed, through a source retrieved in this session. Identify the verification step.
- **Tier 3 — Model background knowledge.** The claim appears consistent with model training knowledge but has not been confirmed against a Tier 1 or Tier 2 source. Tier 3 is never a valid verification source.

4. **Verify Tier 1 items.** For every claim or quotation classified as Tier 1:
- Locate the specific text in the provided document.
- Confirm that the quotation matches the source text exactly.
- Confirm that the section reference, page, or paragraph is accurate.
- Confirm that the claim accurately represents what the source says (not overstated, not understated).
- Mark as: **Verified — Tier 1**, with a note identifying the source document and location.
- If the claim cannot be located in the provided document, reclassify as Unverified and note the discrepancy.

5. **Assess Tier 2 items.** For every claim classified as Tier 2 (independently researched):
- Identify whether the research step and verification are documented in the draft or the underlying workflow.
- If documented and confirmable in this session, mark as **Verified — Tier 2**, with a note identifying the verification step.
- If not documented or not confirmable in this session, mark as **Unverified — requires independent confirmation** and list as an unresolved item.

6. **Flag all Tier 3 items.** For every claim that relies only on model background knowledge:
- Mark as **Unverified — Tier 3 (model knowledge only); must be independently confirmed before reliance**.
- Do not treat apparent consistency with model knowledge as verification.
- List every Tier 3 item as an unresolved item requiring independent research.

7. **Identify unverifiable items.** Flag items that cannot be verified within this session and whose verification requires access to external legal databases, official sources, or documents not provided:
- Case citations where the case text was not provided.
- Statutory text where the current version was not provided.
- Regulatory text where the regulation was not provided.
- Quotations from sources not provided.
Mark each as **Unverifiable in this session — must be confirmed by attorney or researcher**.

8. **Check for incomplete and malformed citations.** For every citation in the inventory, check:
- Is the citation complete? (Case: name, reporter, volume, page, court, year. Statute: instrument, section, version. Regulation: title, part, section, version.)
- Are there missing elements (for example, a case name without a reporter cite, or a statute without a section number)?
- Are there obvious formatting errors (for example, mismatched parenthetical dates or transposed numbers)?
Mark incomplete citations as defects and list them as unresolved items.

9. **Check date claims.** For every date in the inventory:
- Identify whether it is a user-provided date, a date from a provided document, or an asserted date with no source.
- Flag any date that appears to have been computed or derived (rather than quoted from a source) as `[deadline verification required]`.
- Flag any effective date or filing date that has not been confirmed against the applicable source.

10. **Compile unresolved items.** Build a consolidated list of all items marked Unverified, Unverifiable, or defective. For each, state: what the item is, why it is unresolved, and what step is needed to resolve it.

11. **Assemble the validation report** using the output format below. Label it as a draft for attorney review.

## Output Format

Deliver a Source Validation Report with the following sections:

1. **Report Header** — Draft title or description; date of validation; count of items (total | verified | unverified | unverifiable | defective); overall status (Cleared for attorney review / Unresolved items present — see below).
2. **Scope and Limitations** — Whether source documents were provided; any limitations on what could be verified in this session; explicit statement that model background knowledge was not used as a verification source.
3. **Claim Table** — One row per material claim:

| # | Claim | Claim type | Source cited or needed | Support status | Revision needed |
|---|-------|------------|------------------------|----------------|-----------------|

4. **Source Reference Table** — One row per source:

| Source | Source type | Provided by | Claims supported | Limits / notes |
|--------|-------------|-------------|------------------|----------------|

5. **Source Validation Table** — One row per enumerated item:

| # | Item Type | Claim / Citation as Stated | Asserted Source | Source Tier | Status | Notes |
|---|-----------|---------------------------|-----------------|-------------|--------|-------|

Status values: Verified — Tier 1 | Verified — Tier 2 | Unverified | Unverifiable in this session | Defective (incomplete / malformed)

6. **Unsupported Claims List** — Every unsupported or insufficiently supported claim, with required action.
7. **Contradictions List** — Every claim contradicted by available source material.
8. **Recommended Revisions** — Safer wording, placeholders, or deletion recommendations for unresolved claims.
9. **Unresolved Items List** — A numbered list of all unverified, unverifiable, and defective items, each stating: what needs to be checked, by what method, and who is responsible (attorney / legal researcher / client).
10. **Date and Deadline Flags** — A separate list of all date claims flagged `[deadline verification required]` or lacking a confirmed source.
11. **Overall Status Statement** — Either: "No unresolved items. Draft cleared for attorney review, subject to attorney verification checklist." Or: "Unresolved items present. Draft must not be relied upon until all unresolved items are confirmed."

Label the report: **Draft legal work product for attorney review. Not legal advice.**

## Attorney Verification Checklist

- [ ] Every Tier 1 (user-provided document) verification has been confirmed: the attorney has checked that the quotation and claim match the source document.
- [ ] Every Tier 2 (independently researched) item has been confirmed: the research step is documented and the authority has been independently verified.
- [ ] Every item marked Unverified or Unverifiable has been independently checked against an authoritative source by the attorney or a qualified researcher.
- [ ] Every case citation has been confirmed to exist, to be correctly cited, and to stand for the proposition the draft attributes to it.
- [ ] Every statute and regulation citation has been confirmed in the current version applicable to the matter `[deadline verification required]`.
- [ ] Every quotation has been checked against its source text and confirmed to be accurate and in context.
- [ ] Every defective (incomplete or malformed) citation has been corrected or removed.
- [ ] No date in the draft was computed by the agent; all dates are confirmed against source documents or attorney knowledge `[deadline verification required]`.
- [ ] No authority has been accepted as verified on the basis of model background knowledge alone.
- [ ] All items on the Unresolved Items List have been resolved and documented.
- [ ] The validation report and any corrections are retained as part of the matter file.
- [ ] The draft is appropriately labeled as draft legal work product for attorney review, not legal advice.

=== END AGENTCOUNSEL MODULE: source-validation-workflow ===

You are assisting with a legal task using AgentCounsel, a platform-agnostic legal skills library. Use the skill package provided below and follow it exactly.

Operating rules (these always apply):
- Produce draft legal work product for review by a licensed attorney. This is not legal advice and not a final answer.
- Never invent legal authority, citations, quotations, facts, or deadlines. Mark every gap with a visible placeholder such as [CONFIRM: ...] or [VERIFY: ...].
- Identify jurisdiction, governing law, posture, and the relevant date — or flag them as unknown. Never compute a deadline.
- Keep facts, assumptions, analysis, strategy, and verification items visibly separate.
- Follow the skill's Workflow and Output Format. Complete its Attorney Verification Checklist.
- If a Required Input is missing, stop and ask for it. Do not guess.
- This skill has a typed execution contract. Choose a declared mode, evaluate every module activation exactly, and treat missing activation inputs as unresolved rather than loading all modules.

=== BEGIN AGENTCOUNSEL EXECUTION PACKAGE: Legal Research Memo ===
AgentCounsel selective execution package for Legal Research Memo.

This portable package contains every possible resource. For true prompt-size reduction, use the MCP get_skill_context tool or another client that sends only the selected bundle.

# Legal Research Memo

## Purpose

## Use When

## Required Inputs

If the legal question, jurisdiction, or relevant facts are not provided, stop and request them. Do not begin analysis by guessing. Do not fabricate facts to make the question answerable.

## Do Not Use When

## Legal Safety Rules

## Workflow

10. **Assemble the memo** using `templates/legal-research-memo.md` and label it as a draft for attorney review.

## Output Format

Deliver a complete memo using the structure in `templates/legal-research-memo.md`:

Use `[CONFIRM: ...]` wherever a fact, authority, or conclusion is unverified. Do not omit a section because it is difficult to fill; instead, mark it with a placeholder and a note.

## Attorney Verification Checklist

=== END AGENTCOUNSEL CORE SKILL ===

=== END AGENTCOUNSEL SPEC ===

# Legal Research Memo

> Draft legal work product for attorney review. Not legal advice.
> Do not paste client-sensitive facts into a reusable copy of this template.

---

**PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT**

---

## Question(s) Presented

*State each legal question precisely: who did what, under which legal framework, in which jurisdiction, and at what time. Limit each question to one discrete legal issue.*

1. [Question 1: e.g., "Under [governing law], did [party] [action] so as to [legal consequence] as of [date]?"]
2. [Question 2, if applicable]
3. [Add as needed]

---

## Brief Answer(s)

---

## Facts

---

## Assumptions

---

## Discussion / Analysis

---

### Issue 1: [Restate the first legal sub-question]

**Rule**

[State the applicable legal rule or standard. Identify the source precisely: case name, court, year, and specific holding; or statute/regulation with section number and version. Example: "Under [Statute], § [X], a party must [requirement]. [Citation / source]. If the authority cannot be verified in this session, write: `[CONFIRM: rule — no verified authority available for this proposition; attorney must supply and verify before relying on this analysis.]`"]

**Application**

**Conclusion on Issue 1**

---

### Issue 2: [Restate the second legal sub-question, if applicable]

**Rule**

[Rule with citation / source, or `[CONFIRM]` placeholder]

**Application**

[Application to facts, with element-by-element analysis where applicable]

**Conclusion on Issue 2**

[Tentative conclusion with confidence level]

---

*[Add additional Issue sections as needed.]*

---

### Adverse Authority and Counterarguments

---

## Conclusion

[Question 1]: [Summary answer with confidence level and key dependencies.]

[Question 2, if applicable]: [Summary answer with confidence level and key dependencies.]

**Key variables:** [List the two to four facts or legal determinations that most affect the outcome. Attorney should focus verification effort here first.]

---

## Authorities Cited

---

## Open Items / Attorney Verification

---

=== END AGENTCOUNSEL MODULE: legal-research-memo-template ===

> Shared reference material supporting the AgentCounsel legal-research and legal-methodology skills, used to help produce draft legal work product for attorney review — not legal advice.

# Citation Type Taxonomy

This reference gives a reviewer a vocabulary for classifying what *kind* of
citation a passage in a draft actually is, before deciding what verifying it
would require. Different citation forms carry different verification needs
and different hallucination risks — an `id.` is only meaningful in relation
to the citation before it, while a full case citation can be checked in
isolation. This taxonomy does not verify anything itself; it tells a skill
which question to ask for each citation form it finds.

---

## How Skills Use This File

---

## Citation Types

### Full case citation

### Short-form case citation

### Supra citation

### Id. citation

**What it looks like:** A citation of `Id.` or `Id. at [page]`, referring to
the same source as the immediately preceding citation.

### Statutory citation

**What it looks like:** A citation to a statute or code section, e.g. a
pattern like `15 U.S.C. § 78j(b)` or a state-code citation, typically title
or chapter, section symbol, and section number.

### Regulation citation

**What it looks like:** A citation to a codified administrative regulation,
e.g. a pattern like `17 C.F.R. § 240.10b-5`, or a citation to an agency's
Federal Register document by volume and page.

### Law-journal citation

**What it looks like:** A citation to a law review or journal article, e.g.
a pattern like `Author Name, Article Title, 100 Some L. Rev. 1 (2020)`.

### Unknown / partial citation

---

## Verification Workflow Per Type

---

## Reviewer Notes

=== END AGENTCOUNSEL MODULE: citation-type-taxonomy ===

# CourtListener

> Reference material supporting the AgentCounsel skill library, used to help produce draft legal work product for attorney review — not legal advice.

## 1. Source

This connector documents the **free, mostly no-key surface**. CourtListener also exposes PACER-derived data and paid features; both are out of scope here.

## 2. In scope — what CourtListener can verify

CourtListener's corpus and the surface this connector documents support verifying:

## 3. Out of scope — what CourtListener does not cover

A skill must keep its existing placeholder discipline for anything in this list:

If a skill needs to verify any of the above, treat the connector as unavailable for that placeholder and retain the existing `[VERIFY-CITE: ...]` / `[CONFIRM: ...]` flag.

## 4. Surface — where to hit

### 4a. Public web search (no API)

The simplest surface, available in any browser or HTTP-capable tool:

- **Search URL pattern:** `https://www.courtlistener.com/?q=<URL-encoded-query>`
- **Opinion page URL pattern:** `https://www.courtlistener.com/opinion/<opinion-id>/<slug>/`

Use this surface for: quick visual confirmation of a case, sharing a stable URL with the attorney, or environments without HTTP tooling.

### 4b. Citation Lookup API — the primary verification surface

The most reliable way to resolve a citation string to a specific opinion:

### 4c. Search API

For richer queries (court filter, date range, case-name search):

Use this for: locating a case by partial information when only a name or topic is known, narrowing within a specific court or date window.

### 4d. Opinion fetch

For pulling the full text of a verified opinion:

### 4e. MCP tool surface

When the user's environment includes an MCP server that wraps CourtListener (community wrappers exist; check the user's installed MCP servers), prefer the MCP tool surface over raw URL fetches — it handles authentication, retries, and rate-limit backoff. The connector contract is the same: query, record the verified URL, fall back to placeholder if unavailable. **This connector does not configure or install any MCP server** — see `connectors/README.md`.

## 5. Calling pattern from a skill

When a skill's Workflow reaches a citation it would otherwise mark `[VERIFY-CITE: ...]` or `[CONFIRM: ...]`, and the citation is in scope under Section 2:

## 6. Fallback behavior

## 7. Limits and known failure modes

## 8. What this connector does not do

All of those remain attorney-verification items. The connector closes the "does this case exist with this citation" question. Nothing more.

=== END AGENTCOUNSEL MODULE: courtlistener-connector-guide ===

# Negative Treatment Check

## Purpose

Organize a disciplined "is this still good law" check for every authority cited in a draft or research memo. For each authority, the skill records the citation as provided, structures a verification plan using available connectors and any user-supplied citator reports, and classifies the authority's verification status. It never asserts from model memory that a case is good law, has been overruled, superseded, limited, or distinguished — model-memory treatment claims are exactly the hallucination this library exists to prevent. Every treatment conclusion is either drawn from a provided, attributed source or left as an explicit `[Verify current law]` item for attorney confirmation. This is draft legal work product for attorney review, not legal advice.

## Use When

## Required Inputs

If no citations are provided, stop and request them. Do not generate or assume citations to check.

## Do Not Use When

## Legal Safety Rules

## Workflow

## Output Format

Deliver the following, in order, labeled `DRAFT — For Attorney Review — Treatment Not Confirmed From Model Knowledge`:

Use `[Verify current law]` on every status not resting on an attorney's own confirmation. Do not fill gaps with invented content.

## Attorney Verification Checklist

=== END AGENTCOUNSEL MODULE: negative-treatment-workflow ===

=== BEGIN AGENTCOUNSEL MODULE: source-validation-workflow ===
Kind: quality-check
Canonical path: skills/legal-methodology/source-validation/SKILL.md
Load when: standard or deep-review mode

# Source Validation

## Purpose

Apply a systematic method for checking that every cited source, authority, quotation, date, and factual claim in a draft legal work product actually exists and says what is claimed. The skill enumerates every citation and claim in the draft, classifies each by source tier using the hierarchy in `core/source-and-citation-discipline.md`, and marks each as verified, unverified, or unverifiable. It produces a source-by-source validation table and a list of unresolved items that must be checked before the draft is relied upon.

## Use When

**Choosing among the verification skills:**

## Required Inputs

## Do Not Use When

## Legal Safety Rules

## Workflow

Assign each item a unique number for tracking. If the draft is long, work section by section and flag the section with each entry.

11. **Assemble the validation report** using the output format below. Label it as a draft for attorney review.

## Output Format

Deliver a Source Validation Report with the following sections:

| # | Claim | Claim type | Source cited or needed | Support status | Revision needed |
|---|-------|------------|------------------------|----------------|-----------------|

4. **Source Reference Table** — One row per source:

| Source | Source type | Provided by | Claims supported | Limits / notes |
|--------|-------------|-------------|------------------|----------------|

5. **Source Validation Table** — One row per enumerated item:

| # | Item Type | Claim / Citation as Stated | Asserted Source | Source Tier | Status | Notes |
|---|-----------|---------------------------|-----------------|-------------|--------|-------|

Status values: Verified — Tier 1 | Verified — Tier 2 | Unverified | Unverifiable in this session | Defective (incomplete / malformed)

Label the report: **Draft legal work product for attorney review. Not legal advice.**

## Attorney Verification Checklist

=== END AGENTCOUNSEL MODULE: source-validation-workflow ===
=== END AGENTCOUNSEL EXECUTION PACKAGE ===

First, confirm which Required Inputs you have and ask me for any that are missing. Then proceed with the Workflow.