From cocounsel-legal
Setup interview for the CoCounsel Legal plugin — captures research defaults, jurisdiction preferences, subscription tier, citation format, and cross-plugin routing. Use on a fresh install, when the user wants to redo their research profile, or to re-check MCP connector status. Runs without a populated CLAUDE.md.
How this skill is triggered — by the user, by Claude, or both
Slash command
/cocounsel-legal:cold-start-interview [--redo | --check-integrations][--redo | --check-integrations]This skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
1. Check `~/.claude/plugins/config/claude-for-legal/cocounsel-legal/CLAUDE.md`. If already populated and no `--redo`, ask before overwriting.
~/.claude/plugins/config/claude-for-legal/cocounsel-legal/CLAUDE.md. If already populated and no --redo, ask before overwriting.[PLACEHOLDER].~/.claude/plugins/cache/claude-for-legal/cocounsel-legal/*/CLAUDE.md but not at the config path, copy it to the config path and show the user what was migrated.~/.claude/plugins/config/claude-for-legal/cocounsel-legal/CLAUDE.md. Date the footer.--redo — re-run the full interview and overwrite the config.--check-integrations — re-scan MCP connectors and refresh the ## Available integrations table without re-running the full interview.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.
Every research query, every citation check, every brief draft reads from the research profile. If defaults aren't captured, every skill asks the same setup questions from scratch. This interview fills the profile once.
Unlike practice-area plugins (litigation, commercial, etc.), this plugin's setup is lighter — it configures research behavior, not a practice framework. The interview takes ~5 minutes, not 10–15.
Read ~/.claude/plugins/config/claude-for-legal/cocounsel-legal/CLAUDE.md:
<!-- SETUP PAUSED AT: --> → greet the user and offer to resume from that section.[PLACEHOLDER] in critical fields) → greet with a summary of current config. Offer --redo or --check-integrations.Ask:
"Before we set up CoCounsel Legal, one question: are you a lawyer or legal professional, or are you a non-lawyer who works with an attorney?"
Capture role. If non-lawyer, ask for attorney contact.
Test the CoCounsel Legal MCP server:
legal_research_start_deep_research with a minimal test query (e.g., "test connectivity") and immediately note whether the server responds (even with an error about the query — a response means connectivity works)."The CoCounsel Legal MCP server isn't connected. You'll need a CoCounsel Legal subscription with the MCP connector enabled. Contact [email protected] to set it up, then come back and run this setup again."
Check optional integrations:
Check ~/.claude/plugins/config/claude-for-legal/company-profile.md:
company-profile.md."Which jurisdictions do you research most often? I can set up to 3 as defaults so you don't have to specify them every time. Examples: California, New York, Federal (9th Circuit)."
If the user names more than 3, note the limit and ask them to pick their top 3.
"What practice areas do you focus on? This helps me route research to the right follow-up tools. For example: employment law, IP litigation, commercial contracts, regulatory compliance."
"When I deliver a Deep Research report, would you prefer:
- Verbatim — the raw Westlaw report, exactly as delivered (default)
- Memo — automatically reformatted as a research memo
- Brief section — automatically reformatted for insertion into a brief
- Executive summary — one-page overview
You can always ask for a different format after delivery. This just sets the default."
"For brief sections and memos, which citation format?
- Bluebook (default)
- ALWD
- Your firm's house style — if so, tell me what's different from Bluebook."
If house style, capture the deviations.
"Do you know your CoCounsel Legal subscription tier? This helps the quota tracker give you accurate usage reports. If you're not sure, I'll leave it blank and you can fill it in later.
- Plan name: standard / professional / enterprise / don't know
- Daily query limit: (e.g., 50)
- Monthly query limit: (e.g., 500)
- Max concurrent sessions: (e.g., 3)"
If the user doesn't know, leave as [PLACEHOLDER] — quota-tracker will still work, it just won't show limits.
"When should I warn you about quota? Default is 80% of monthly limit. Want a different threshold?"
"CoCounsel Legal integrates with other practice-area plugins. I'll check which ones are installed and set up routing."
Scan for installed practice-area plugins by checking for CLAUDE.md files in sibling directories:
litigation-legalcommercial-legalregulatory-legalip-legalemployment-legalcorporate-legalprivacy-legalFor each installed plugin, add it to the routing table. For uninstalled plugins, note it as unavailable.
"Do you have any citations you want to monitor for negative treatment? The authority-tracker agent can check them on a schedule and alert you if a case you rely on gets distinguished, overruled, or otherwise weakened.
You can add these now or later. Just paste citations in any format."
If the user provides citations, add them to the watch list.
Assemble the captured data into the CLAUDE.md template structure:
[PLACEHOLDER] fields with captured values[not configured] (not [PLACEHOLDER] — distinguishes "asked and skipped" from "never asked")*Written by cold-start on YYYY-MM-DD.*~/.claude/plugins/config/claude-for-legal/cocounsel-legal/CLAUDE.mdIf the user wants changes, make them and rewrite. Then confirm:
"Setup complete. Your research defaults are saved. Every CoCounsel Legal skill will read these preferences automatically. You can edit them anytime at
~/.claude/plugins/config/claude-for-legal/cocounsel-legal/CLAUDE.mdor re-run/cocounsel-legal:cold-start-interview --redo."
npx claudepluginhub chgreer1070/claude-for-legal --plugin cocounsel-legalGuides completion of development work by verifying tests, detecting environment, and presenting structured options for merge, PR, or cleanup.
Guides creation and editing of skills using test-driven development with pressure scenarios and subagents to verify agent compliance.
Dispatches multiple subagents concurrently for independent tasks without shared state. Use when facing 2+ unrelated failures or subsystems that can be investigated in parallel.