メニュー
Update your REST-AP catalog to describe your role, expertise, and status to other agents and the manager.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Communicate with other agents in your team — send messages, delegate tasks, list agents, check news. Use when asked to contact another agent or coordinate work.
Programmatically manage an ID Agents team — add/remove agents, sync configs, rebuild and restart the manager, dispatch work to agents and poll replies. Use whenever you edit a team YAML, hit "Manager did not start in time", need to /sync, /deploy, or /agents rebuild a team, or want to talk to or ask an agent.
Build an id-agents team correctly from scratch. Covers YAML structure, per-agent workspaces, role files, library agent templates, skill bundling, and the gotchas (silent default drops, symlinks vs copies, cpSync collisions). Use whenever you are designing a new team config, adding agents to an existing team, or debugging why an agent ended up with a generic personality.
Your agent identity — name and team. Always loaded so you know who you are.
Analyze and choose names based on the perceptual qualities that letters and sounds carry. Use when generating, screening, or defending brand names. Covers the English letter→quality table, compound multiplier effect, rhythm patterns, palindromes, and processing fluency. Derived from Lexicon Branding's research into linguistic sound symbolism.
Required lifecycle for non-trivial work. Create a task, claim it, do the work, mark done. Include the task name in your reply.
| name | catalog |
| description | Update your REST-AP catalog to describe your role, expertise, and status to other agents and the manager. |
| allowed-tools | Bash |
You can update your own catalog to describe what you do, your role, skills, and current status. This information is visible to other agents and the manager via your /.well-known/restap.json endpoint.
curl -s http://localhost:$ID_AGENT_PORT/catalog | jq
curl -s -X PATCH http://localhost:$ID_AGENT_PORT/catalog \
-H "Content-Type: application/json" \
-d '{
"description": "I specialize in TypeScript and React development",
"role": "developer",
"expertise": ["typescript", "react", "node", "testing"],
"status": "available",
"currentTask": "Working on user authentication",
"model": "claude-opus-4-7",
"workingDirectory": "/Users/nxt3d/projects/id2/id-agents/workspace/agents/agents",
"costTier": "high",
"notSuitableFor": ["bulk data crunching", "long-running batch jobs"]
}'
| Field | Description | Example |
|---|---|---|
description | What you do | "Full-stack developer focusing on React" |
role | Your assigned role | "developer", "researcher", "pm" |
expertise | Array of skills | ["typescript", "react", "testing"] |
status | Availability | "available", "busy", "offline" |
currentTask | What you're working on | "Implementing login flow" |
model | Underlying LLM model id | "claude-opus-4-7", "claude-sonnet-4-6", "claude-haiku-4-5-20251001" |
workingDirectory | Absolute path to your agent workspace | "/Users/alice/projects/foo/workspace/agents/coder" |
costTier | Relative cost of routing work to you — low, medium, or high | "high" (Opus), "medium" (Sonnet), "low" (Haiku) |
notSuitableFor | Array of work patterns where the manager should route elsewhere | ["bulk data crunching", "image generation", "production deploys"] |
Update your catalog when starting work (set status to "busy") and when done (set to "available"). Keep model, workingDirectory, costTier, and notSuitableFor accurate so the manager can route work to the right agent: costTier and notSuitableFor together act as routing hints, while model and workingDirectory let other agents reason about your capabilities and where your artifacts live.