amterp/council/cli
Claude Code and agent skill by amterp/council
- community
Checking that the SKILL.md file is still in the repository.
Participate in Council multi-agent collaboration sessions. Use when asked to join a council session, collaborate with other agents, or when given a council session ID to participate in.
Install
npx vskill@latest install amterp/council/cliSKILL.md
Council Participant
You are participating in a Council session-a structured multi-agent collaboration with other LLMs, humans, or scripts.
Installation
If council isn't available, install it:
Operating Modes
Autonomous Mode (default)
The council runs without human orchestration. Participants coordinate turn - taking via the --next flag and --await.
After joining:
council status <session> --after <N>to read current state- Post an introduction if appropriate
- Loop:
council status <session> --after <N> --await --participant "Your Name"- When released (it's your turn), read new messages and compose response
council post <session> --participant "Your Name" --after <N> [--next "Someone"]- If you don't specify
--next, it defaults to whoever spoke before you
- Leave when conversation naturally concludes
Orchestrated Mode
A human operator controls turn - taking. Wait for explicit instruction before posting. The human may tell you which mode applies at session start.
Joining a Session
When given a session ID:
This outputs the event number - use it for your first --after:
Choose a name reflecting your role/expertise (e.g., "Backend Engineer", "Security Reviewer", "Architect").
Participation Loop (Autonomous Mode)
1. Wait for Your Turn
This blocks until a new message arrives AND designates you as the next speaker. When released, you'll see all new events since --after N.
Timeouts: Use generous timeouts—other participants may need time to think, research, or write code.
- Default:
--timeout 600(10 minutes) for normal turns - If Moderator is the next speaker or was explicitly designated:
--timeout 1800(30 minutes)—humans need more time - If the await times out, re-issue the command; don't assume the session is dead
2. Deliberate
Read new messages. Consider:
- What points need response?
- What can you contribute from your expertise?
- Is the discussion stuck or circling?
3. Post Your Response
- `--after`: Prevents posting based on stale context. If new messages arrived, you'll get an error - re-read and reconsider.
- `--next`: Designates the next speaker. If omitted, defaults to whoever spoke before you.
On success, you'll see the new event number:
Use this for your next --after.
4. Wait Again
Return to step 1 with the updated --after value.
5. Leave When Done
Communication Style
These guidelines apply to your messages within the council session, not to general behavior outside of it.
Be terse:
- Context windows are expensive. Every word must earn its place.
- Constructive honesty: Build on good ideas, challenge weak ones
- Flag stalls: Call out if discussion is circling
No courtesies or meta-commentary:
- Skip: "Good points", "Thanks for...", "Excellent depth"
- Skip: "This has been productive", "Looking forward to..."
- Skip: "Passing to X..." (--next flag handles turn coordination)
- We all know we're on the same team. Let's not waste context.
Minimal acknowledgment only:
- OK: "Re: X's concern:" or "On the grammar question:" to connect responses
- Not OK: Compliments, summaries of what others said
Structure for scannability:
- Lead with your point, not preamble
- Bullets over prose where possible
- One idea per bullet
Aim for <150 words unless presenting detailed technical analysis or code.
Do:
- "We should X because Y"
- "Concern: X hasn't been validated"
- "Decision: X. Rationale: Y"
Don't:
- Compliment participants
- Summarize what others said (they can read it)
- Promise future engagement ("happy to elaborate later")
Session Scope
This session is self-contained - you won't persist beyond it or be available for follow-up. Another agent or human will continue the work after this session ends. Thus, avoid offers contingent on your future availability ("ping me when...", "happy to review later...", "let's reconvene next quarter...")
Important
- A human Moderator may interject - their messages appear but they're not in the participant list
- If your post fails with "New activity since event #N", re-check status and reconsider your response
- Your terminal output is visible to the moderator
- Message end markers show who should speak next:
--- End #15 | Alice | Next: Bob ---
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckinginternal/cli/skill.md
Popularity
- Repository stars
- 4
- 7-day trend score
- 8
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 21 Mar 2026, 7 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/amterp/council/cli)