Launch Review

/launch-review

1. Load ~/.claude/plugins/config/claude-for-legal/product-legal/CLAUDE.md → framework + calibration. Stop if placeholders.
2. Get PRD + related docs. If tracker connected, pull ticket and comments.
3. Walk every framework category using the workflow below.
4. Calibrate each finding against the table. Novel = flag explicitly.
5. Output review memo in house format. Post summary to ticket if connected.
6. Hand off: marketing-claims-review if substantial marketing; feature-risk-assessment if a finding needs depth.

/product-legal:launch-review PROJ-1234

Matter context

Matter context. Check ## Matter workspaces in the practice-level CLAUDE.md. If Enabled is ✗ (the default for in-house users), skip the rest of this paragraph — skills use practice-level context and the matter machinery is invisible. If enabled and there is no active matter, ask: "Which matter is this for? Run /product-legal:matter-workspace switch <slug> or say practice-level." Load the active matter's matter.md for matter-specific context and overrides. Write outputs to the matter folder at ~/.claude/plugins/config/claude-for-legal/product-legal/matters/<matter-slug>/. Never read another matter's files unless Cross-matter context is on.

Destination check

Before producing output, check where it's going. If the user has named a destination (a channel, a distribution list, a counterparty, "everyone"), ask whether it's inside the privilege circle. Public channels, company-wide lists, counterparty/opposing counsel, vendors, and clients (for work product) waive the protection. When the destination looks outside the circle, flag it and offer (a) the privileged version for legal only, (b) a sanitized version for the broader channel, or (c) both — don't silently apply a privileged header and then help paste it somewhere the header won't protect it. See the canonical ## Shared guardrails → Destination check in this plugin's CLAUDE.md.

Purpose

Read the PRD, check every category in this team's framework, calibrate against what actually blocks here (per ~/.claude/plugins/config/claude-for-legal/product-legal/CLAUDE.md), and output a review in house format. Goal: a PM reads it and knows exactly what has to happen before they ship.

Load calibration

Read ~/.claude/plugins/config/claude-for-legal/product-legal/CLAUDE.md:

• ## Review framework — the categories to check
• ## Risk calibration — what blocks vs. what's FYI at this company
• ## Launch review process — output format
• ## Escalation — when to route up

The calibration table is the difference between this skill and a generic checklist. If the table says "new data collection → PIA, ships in 1-2 days," don't write "this might require a full DPIA and regulatory consultation." Match the team's actual practice.

Workflow

Step 1: Get the inputs

• PRD — from file, Drive, or the launch tracker ticket
• Spec/design doc — if separate
• Marketing plan — if there is one (hands off to marketing-claims-review if substantial)
• Launch date — for urgency calibration
• Launch tracker ticket — if connected, pull it for context and comments

If Jira/Linear MCP is connected, pull the ticket history — often there's context in earlier comments that the PRD doesn't capture.

Step 2: Understand what's launching

Before the checklist, answer in plain English:

• What does this thing do?
• Who uses it — existing users, new users, a new segment?
• What's new vs. what's an extension of something already reviewed?
• Any new data, new vendors, new claims, new jurisdictions?

AI detection — run before the framework walk. Check whether this launch uses AI in any form: a third-party model, an internally built model, an AI-powered vendor feature, automated scoring or classification, generative content, recommendations, predictions. Look for this even if the PRD doesn't label it "AI" — words like "intelligent", "automated", "personalized", "generated", "suggested" are tells.

If AI component detected → flag it, then run /ai-governance-legal:use-case-triage [feature] alongside the framework walk. Category 8 below handles the detail; this flag ensures it's never skipped even if the PRD is vague.

Step 3: Walk the framework

For each category in ~/.claude/plugins/config/claude-for-legal/product-legal/CLAUDE.md → Review framework. If the team doesn't have one, use the 8-category default below. The categories are stable framing concepts; within each category, research the regulatory regimes applicable to the product's sector, audience, and jurisdictions before calibrating severity. What blocks in one jurisdiction or sector may be routine in another — ~/.claude/plugins/config/claude-for-legal/product-legal/CLAUDE.md captures the team's calibration.

No silent supplement. If a research query to the configured legal research tool (Westlaw, CourtListener, regulator sites, or firm platform) returns few or no results for a regime, enforcement precedent, or regulator guidance, report what was found and stop. Do NOT fill the gap from web search or model knowledge without asking. Say: "The search returned [N] results from [tool]. Coverage appears thin for [regime / topic]. Options: (1) broaden the search query, (2) try a different research tool, (3) search the web — results will be tagged [web search — verify] and should be checked against the issuing authority before relying, or (4) flag as unverified and stop. Which would you like?" A lawyer decides whether to accept lower-confidence sources. Source attribution tiering. Tag every citation in the review with its source. For model-knowledge citations, use one of three tiers rather than a single blanket "verify" tag: - [settled] — stable, well-known statutory and regulatory references unlikely to have changed (e.g., FTC Act § 5, GDPR Art. 33, CCPA § 1798.100). Still verify before relying on it to clear a launch, but lower priority. - [verify] — model-knowledge citations that are real but should be verified: specific implementing regulations, agency guidance, enforcement actions, case holdings, thresholds, effective dates, post-2023 amendments. - [verify-pinpoint] — pinpoint citations (specific subsection letters, volume/page numbers, paragraph numbers) carry the highest fabrication risk and should ALWAYS be verified against a primary source. Tool-retrieved citations keep their source tag ([Westlaw], [CourtListener], [regulator site], or the MCP tool name); web-search citations remain [web search — verify]; user-supplied citations (from the PRD or seed materials) remain [user provided]. The tiering surfaces the real verification work — a reader who verifies everything verifies nothing. Never strip or collapse the tags.