Run the cold-start interview to learn your commercial contracts practice and write your team practice profile. Use on first use of the plugin, when `~/.claude/plugins/config/claude-for-legal/commercial-legal/CLAUDE.md` is missing or still contains template placeholders, or when the user says "set up

/cold start interview 1. Check ~/.claude/plugins/config/claude for legal/ai governance legal/CLAUDE.md — if populated and no redo , confirm before overwriting. 2. Run the interview using the workflow below (includes Part 0 role + integration check). 3. Seed docs: AI/acceptable use policy (URL or file), a prior impact assessment, key vendor AI agreements, model inventory or allowlist/blocklist if they exist. Read all provided. 4. Extract: policy commitments and prohibitions, vendor positions (note gaps vs. stated), impact assessment structure, approved/prohibited tool lists. 5. Migration: if a populated CLAUDE.md (no [PLACEHOLDER] markers) exists at ~/.claude/plugins/cache/claude for legal/ai governance legal/ /CLAUDE.md but not at the config path, copy it to the config path and tell the user what was migrated. 6. Write ~/.claude/plugins/config/claude for legal/ai governance legal/CLAUDE.md (create parent directories as needed). Show summary. Offer first task. Flags redo — re run the full interview and overwrite ~/.claude/plugins/config/claude for legal/ai governance legal/CLAUDE.md . check integrations — re scan available MCP connectors and refresh the Available integrations table in ~/.claude/plugins/config/claude for legal/ai governance legal/CLAUDE.md without re running the full interview. Use after setting up a new connector (Slack, document storage, scheduled tasks). When probing: only report ✓ if an MCP tool call actually succeeded. Configured but untested connectors should be marked ⚪ with a one line how to for confirming. Never report ✓ based on .mcp.json declarations alone — that misleads users into thinking something is wired up when it isn't. Purpose Learn how this AI governance team works — what role the company plays in the AI supply chain, which regulations actually apply to them, what their red lines are for AI use cases, and what good impact assessment looks like here. Write it into the plugin config so every other skill reads from the same understanding. AI governance postures vary enormously. A company that builds AI products for enterprise customers has almost nothing in common with a company that deploys off the shelf AI tools internally. The interview figures out which one this is before anything else — because builder obligations and deployer obligations are nearly opposite exercises. Cold start check Read ~/.claude/plugins/config/claude for legal/ai governance legal/CLAUDE.md : Does not exist → start the interview. Contains <! SETUP PAUSED AT: → greet the user and offer to resume from that section. Contains [PLACEHOLDER] markers but no pause comment → the template was never completed; offer to start fresh or resume from wherever the placeholders begin. Populated (no placeholders, no pause comment) → already configured; skip unless redo . The template structure lives at ${CLAUDE PLUGIN ROOT}/CLAUDE.md — use it as the section scaffold. Write the completed practice profile to the config path, creating parent directories as needed. If a CLAUDE.md exists at the old cache path ~/.claude/plugins/cache/claude for legal/ai governance legal/ /CLAUDE.md but not at the config path, copy it forward to the config path before proceeding. Check for the shared company profile Look for ~/.claude/plugins/config/claude for legal/company profile.md . If it exists: Read it. Show a one line confirmation: "You're [name], [practice setting], at [company], [industry], operating in [jurisdictions]. Right? (Or say 'update' to change the shared profile.)" If confirmed, skip the company questions — go straight to the plugin specific ones. If it doesn't exist: You'll be the first plugin this user set up. After the orientation and fork, ask the company questions and write them to the shared profile (per the template at references/company profile template.md in the plugin root), then continue with the plugin specific questions. Tell the user: "I've saved your company profile — the other legal plugins will read it and skip these questions." The company questions that belong in the shared profile (and should NOT be re asked if it exists): practice setting, company name, industry, what you sell, size, jurisdictions, regulators, risk appetite, escalation names. The plugin specific questions (playbook positions, review framework, house style, supervision model, etc.) stay per plugin. Install scope check Before the orientation, if you notice the working directory is inside a project (not the user's home directory), flag it. Say once: Heads up — it looks like this plugin may be project scoped, which means I can only read files in [current directory]. If you'll want me to read documents from elsewhere (Downloads, Documents, Dropbox), install user scoped instead — see QUICKSTART.md. You can continue with project scope, but you'll need to move files into this folder. Ask the user to confirm before proceeding: continue with project scope, or pause to reinstall user scoped. If the working directory is the user's home directory, skip this check silently. Before the interview starts Open with the fork first preamble. Keep it to 3 4 short lines. Ask quick or full before anything else. ai governance legal is for people who run AI governance: use case triage, impact assessments, vendor AI review, policy monitoring. Not your area? /legal builder hub:related skills surfacer . 2 minutes gets you your role, practice setting, and which AI regulatory regimes apply (EU AI Act, NIST, state AI laws), plus working defaults for use case triage thresholds, AIA format, and vendor AI positions. 15 minutes adds your use case registry and red lines, governance tiers, vendor AI playbook positions, escalation matrix, AIA house style template extracted from a seed assessment, and the AI policy commitments extracted from your actual policy. Quick or full? (Upgrade any time with /cold start interview full .) Quick start path: ask only Part 0 (role, practice setting, integrations) and regulatory scope. Write the config with [DEFAULT] markers on everything else. Close with: "Done. You can start using the commands now. I've used sensible defaults for use case triage thresholds, AIA format, and vendor AI positions. When a skill's output feels off, that's usually a default you should tune — it'll tell you which. Run /ai governance legal:cold start interview full anytime to do the whole interview, or /ai governance legal:cold start interview redo <section to re do one part." Full setup path: the existing interview flow below. After the user picks, give the fuller orientation described next, then proceed to Part 0. After the user picks quick or full Give the fuller orientation. One paragraph, in your own voice: "This plugin maintains: your practice profile (governance tiers, red lines, policy commitments), a use case registry, impact assessments, and vendor AI reviews — all in ~/.claude/plugins/config/claude for legal/ai governance legal/ . It learns how you actually work — your practice, your risk calibration, your house conventions — and writes that into a plain text file the plugin reads from every time. Everything you answer can be changed later." Then: "Ready? A few quick questions first, then we'll go deeper." Why this matters (offer if the user pushes back on the time cost). Every triage, impact assessment, vendor review, and policy monitor sweep reads from the configuration this interview writes. A generic configuration gives generic output — a default use case registry, default red lines, a default vendor AI position matrix, and a triage that treats a resume screening tool the same as an expense anomaly flagger. Telling the plugin whether the user is a builder or a deployer, where the red lines are, and what they require from vendors is what makes the difference between "an AI governance AI tool" and "a tool that knows your posture." Fresh professional profile. Setup builds a fresh professional profile from the user's answers and the documents they explicitly share. It does not read the user's personal Claude history, unrelated conversations, or their home directory CLAUDE.md. If something relevant surfaces in the current conversation context (e.g., they mentioned their company earlier), ask before using it — do not fold anything personal into the practice profile unless the user types it or approves it. Corollary: the interview's inputs are the user's typed answers and documents they explicitly share. Do not pull from ambient context, prior sessions, or user memory to fill in gaps. Interview pacing Assume the answer exists somewhere. When a question asks for information that's probably written down somewhere — company description, playbook, escalation matrix, style guide, handbook, jurisdiction list, matter portfolio — prompt for a link or a paste before asking the user to type it from memory. "Paste a link or a doc, or give me the short version" is the default ask for anything that's more than a sentence. An interviewer who makes people re type what they've already written has failed the first job of an interviewer. Batch size — count subparts. "Never ask more than 2 3 questions in one turn" means 2 3 answerable prompts , counting subparts. One question with 5 subparts is 5 questions. The test: can the user answer without scrolling? If the questions don't fit on one screen, it's too many. Prefer structured tap through questions where possible — they don't require scrolling or typing. Pause for real answers. Some questions are quick (pick A/B/C). Others need the user to type, describe, or share a document. When a question needs more than a quick tap: Ask and wait. Say explicitly: "This one needs a typed answer — I'll wait." Do not move to the next question until the user responds. For uploads or shared documents: "Paste the contents, share a file path, or say 'skip for now.' If you skip, I'll flag the gap in your practice profile so you can fill it later." Then actually wait. Before writing the practice profile: review the interview and list any questions that were skipped or answered with placeholders. Say: "Before I write your configuration, here's what's still open: [list]. Want to fill any of these now, or leave them as placeholders?" Then wait for the answer. Never write a practice profile with silent gaps. Every placeholder should be a deliberate choice the user made to skip, not a question that scrolled past. Pause and resume. Tell the user up front: "If you need to stop, say 'pause' (or 'stop', or 'let me come back to this') and I'll save your progress. Run /ai governance legal:cold start interview again later and I'll pick up where you left off." When the user pauses, write a partial configuration to ~/.claude/plugins/config/claude for legal/ai governance legal/CLAUDE.md with a <! SETUP PAUSED AT: [section name] — run /ai governance legal:cold start interview to resume comment at the top and [PENDING] markers (distinct from [PLACEHOLDER] ) on unanswered fields. When setup re runs and finds a paused config, greet the user: "Welcome back. You paused at [section]. Your earlier answers are saved. Pick up where we left off, or start over?" Do not re ask questions already answered. Verify user stated legal facts as they come up in setup. When the user answers an interview question with a specific rule citation, statute number, case name, deadline, threshold, jurisdiction, or registration number — and it's something you can sanity check — do the check before writing it into the configuration. If what they said conflicts with your understanding or with something they've pasted, surface it: "You said the threshold is X; my understanding is Y — can you confirm which goes in the profile? [premise flagged — verify] " A wrong fact written into CLAUDE.md propagates into every future output; catching it here

cold-start-interview