counsels-write-memo

/counsels-write-memo [--persona ]

Produce a legal, policy, or ethics memo on a specific question.

⚠️ NOT LEGAL ADVICE. This skill produces craft-register writing in the voice of canonical figures. It is a writing tool and a reasoning lens, not a substitute for licensed counsel. Any decision with real legal stakes requires a real attorney admitted to practice in the relevant jurisdiction.

What this does

This skill is the counsel equivalent of /engineers-write-spec — the upstream artifact that downstream work obeys. Before a decision is made, the memo establishes the question, the relevant facts, the applicable framework, the analysis, the conclusion, and the caveats and limits.

The memo serves three downstream consumers:

1. The decision-maker — they read the memo and decide.
2. The reviewer (a peer, a senior, a real lawyer if one is being engaged) — they check the analysis.
3. The future reader — months or years later, they read the memo to understand why a given choice was made.

When to use

• A specific legal, policy, or ethics question is in front of the project and the analysis benefits from being captured in writing.
• Multiple framings are possible and the team needs to commit to one (or to capture the disagreement).
• An ethical question requires reasoning through a framework rather than gut feel.
• A regulatory question needs choice-architecture or behavioral analysis.

Not for: live legal advice on a real matter (the disclaimer is not optional); line-level policy nits (use /counsels-review); craft conversations about a single concept (use /counsels-channel <persona>); auto-generated documentation.

Instructions for Claude

When this skill is invoked with a <question> argument and optional --persona:

1. Resolve the project root the same way /counsels-project-init does. Verify CLAUDE.md (or README.md) exists and the counsel/ directory is present. If counsel/ is missing, recommend running /counsels-project-init first; do not auto-create.

2. Read the project specification:

   • CLAUDE.md — orchestrator-mode notes, current memo slug, conventions
   • README.md — what the project does, who it serves
   • Prior memos at counsel/memos/
   • Prior reviews at counsel/reviews/
   • The actual decision, policy, or practice the question concerns
   • .great-authors/project.md if cross-craft project

3. Resolve the persona to dispatch. If --persona is given, use it. Otherwise auto-select by signal:

   Signal in the question / project	Default persona
   "civil rights", "equal protection", "discrimination", constitutional	ruth-bader-ginsburg-counsel
   "litigation strategy", "long game", "precedent", "test case"	thurgood-marshall-counsel
   "textualist", "originalist", "what does the text say", statutory	antonin-scalia-counsel
   "code as law", "regulation by code", digital, "four modalities"	lawrence-lessig-counsel
   "antitrust", "monopoly", "platform power", "net neutrality"	tim-wu-counsel
   "privacy", "right to be let alone", surveillance, data	louis-brandeis-counsel
   "regulation", "rule-making", "nudge", "choice architecture"	cass-sunstein-counsel
   "totalitarianism", "banality of evil", "action vs behavior"	hannah-arendt-counsel
   "veil of ignorance", "justice as fairness", "original position"	john-rawls-counsel
   None of the above	john-rawls-counsel (the ethical-reasoning default)

   Document the choice in the memo's frontmatter.

4. Dispatch the persona via the Agent tool with subagent_type: "great-counsels:<persona-slug>-counsel". The brief must include:

   • The question (the user's <question> argument, plus any context)
   • All bible files read above (paths only — the persona reads them)
   • The actual decision, policy, or practice the question concerns
   • Prior memos and reviews that inform this question
   • The output target: counsel/memos/<slug>.md
   • The required structure (below)
   • Length target: 600-1,200 words
   • Mandatory disclaimer block at the top of the artifact

5. The output structure:

---
title: <Memo title>
slug: <slug>
persona: <persona-slug>
created: YYYY-MM-DD
status: draft | proposed | accepted | superseded
---

# Counsel memo: <Title>

> ⚠️ **Craft register only — not legal advice.** This memo is a reasoning exercise in the voice of <persona display name>, not a representation by counsel. For matters with real legal stakes, retain an attorney admitted in the relevant jurisdiction.

## Question

<One paragraph. The specific question the memo answers.>

## Relevant facts

<Bulleted list. The facts that bear on the question. Honest about what is known and what is assumed.>

## Applicable framework

<2-3 paragraphs. The legal/policy/ethical framework the persona is applying. RBG would apply equal-protection doctrine; Lessig would apply the four modalities; Rawls would apply justice as fairness. Name the framework and its load-bearing parts.>

## Analysis

<2-4 paragraphs. The actual reasoning — apply the framework to the facts. Show the work. Cite the cases, principles, or arguments the persona would cite.>

## Conclusion

<One paragraph. The persona's verdict, with the reasoning compressed.>

## Caveats and limits

<Bulleted list. What the memo does NOT decide. What further analysis would require. **Always include: this is craft-register reasoning, not legal advice.** What facts, if different, would change the conclusion.>

6. Save the doc to counsel/memos/<slug>.md. If the file exists, ask the user whether to overwrite, save as <slug>-v2.md, or skip.

7. Report:

   📝 Saved to counsel/memos/<slug>.md (<word-count> words, drafted by <persona>).
   ⚠️ Craft register only — not legal advice.
   
   Question:    <one-line summary>
   Conclusion:  <one-line summary>
   
   Next:
   - /counsels-review counsel/memos/<slug>.md to dispatch a panel for review
   - /counsels-channel <other-persona> to refine specific sections (e.g., Scalia on the textualist counter-reading, Wu on the platform-power dimension)
   - For real legal stakes: retain a licensed attorney in the relevant jurisdiction.
   

What the skill does NOT do

• Give legal advice. The persona is a craft channel; the memo is a reasoning exercise. It does not represent the user.
• File a brief. Memos produced by this skill are persona-voice writing exercises. Filing them in a court is not a use case the plugin supports.
• Resolve hard moral calls. Arendt and Rawls personas can reason through ethical frameworks; they cannot tell the user what the right thing to do is for the user's specific situation.
• Verify quotations or citations. Personas reference their subjects' published works as part of voice. Treat any quoted line as suggestive, not authoritative; check the source before relying on it.

Notes

• The memo is a contract on reasoning, not on action. If the conclusion conflicts with the user's intuition, the conflict is information; the next move is a real lawyer or ethicist, not a redraft.
• A project may have multiple memos over time (one per question). Each gets its own slug — set via CLAUDE.md's Current memo: field.
• For veil-of-ignorance analysis specifically, use --persona rawls and request the veil format explicitly. Filed for v1.0 as /counsels-veil-of-ignorance.
• The disclaimer block is not optional. Do not strip the "Craft register only — not legal advice" line from any memo.