菜单
Operate JClaw itself from chat — list/create/update agents, add or test MCP servers, toggle tools/skills on agents, read or write config. Use whenever the user asks to change JClaw's own configuration rather than perform an external task.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Read a business card image, extract contact details (name, phone, email, company, etc.), append them to a Google Sheet, then send a welcome email and WhatsApp message to the contact.
>-
Fetch and summarize YouTube video transcripts using a residential IP proxy to bypass cloud IP blocks.
Create, inspect, and edit Microsoft PowerPoint presentations and PPTX decks with reliable layouts, templates, placeholders, notes, charts, and visual QA.
Perform JIRA operations using the MCP server as primary, with automatic fallback to REST API for operations that don't work via MCP (e.g., story points via Agile estimation endpoint, epic linking via parent field). Includes sprint workflow helpers.
Produce thoughtful design artifacts (slide decks, prototypes, mockups, landing pages, dashboards) using HTML/CSS/JS/SVG as the medium.
| name | jclaw-api |
| description | Operate JClaw itself from chat — list/create/update agents, add or test MCP servers, toggle tools/skills on agents, read or write config. Use whenever the user asks to change JClaw's own configuration rather than perform an external task. |
| version | 0.1.0 |
| tools | ["jclaw_api"] |
| icon | ⚙️ |
Use this skill when the user asks you to change JClaw's own state — adding an MCP server, creating an agent, toggling a tool on an agent, editing a config value, etc.
The jclaw_api tool exposes JClaw's own HTTP API, and the list of callable endpoints is discovered at runtime — there is no fixed catalog to drift. Do not invent endpoints, and never call one that discover does not list. Every /api/ endpoint is callable by default unless it is deny-listed, so coverage is automatic — a newly-added endpoint appears in discover with no skill edit. If discover does not list something, it is deliberately off-limits (see below); to hide a sensitive endpoint, mark its controller action @ChatHidden (or add its prefix to the tool's deny-floor) in the backend, not in this skill.
jclaw_api with action="discover" (optionally pass a filter substring like "agent" or "mcp" to narrow the list). It returns the live catalog: verb, path, a one-line summary, and a body hint for mutating verbs.jclaw_api with:
method: one of GET, POST, PUT, PATCH, DELETEpath: starts with /api/ exactly as discover reports it — never include the host or schemebody: a JSON object matching the operation's request shape (POST/PUT/PATCH only)query: optional, only when an operation documents query parametersThe tool returns HTTP <status>\n<body> so you can read both. A 4xx response usually carries a JSON error field that tells you what to correct.
These staples don't need a discover round-trip first; run action="discover" for anything else.
GET /api/agents (list); POST /api/agents body name, modelProvider, modelId (create); PUT /api/agents/{id} (update); DELETE /api/agents/{id} (delete an agent). The built-in main agent cannot be deleted — DELETE on it returns 409; don't attempt it.POST /api/mcp-servers body name, enabled, transport (STDIO or HTTP), then command/args/env or url/headers; confirm it boots with POST /api/mcp-servers/{id}/test.PUT /api/agents/{id}/tools/{name} with { "enabled": <bool> }.POST /api/agents/{id}/skills/{name}/copy with an empty body (installs and enables); PUT /api/agents/{id}/skills/{name} only toggles an already-installed skill.GET /api/config (all, masked) or GET /api/config/{key}; POST /api/config body { "key": ..., "value": ... }. Confirm with the user before touching provider.*, auth.*, chat.*, or shell.*.Common composition — "create agent X with skill Y":
POST /api/agents → capture the returned id.POST /api/agents/{id}/skills/Y/copy with {} → installs Y into the new agent's workspace and enables it.The following endpoints exist in JClaw's API but are deliberately not callable through jclaw_api, and the tool itself will refuse them (they're also excluded from discover):
POST /api/chat/send / POST /api/chat/stream / POST /api/chat/upload — these would create a recursion loop where this skill sends a chat that itself invokes this skill again. If the user wants to send a chat programmatically, suggest the per-channel webhook endpoints instead.
POST /api/auth/login / POST /api/auth/setup / POST /api/auth/reset-password — credential operations belong in the Settings UI.
All /api/api-tokens routes — minting or revoking API tokens through chat would be a privilege-escalation surface.
/api/webhooks/* — verified by their own signature mechanisms; not for in-process use.
/api/events — Server-Sent Events; the tool buffers full responses and isn't suited to streams.
/api/bindings, /api/channels/telegram/* — channel routing and Telegram bot tokens; editing these could redirect or impersonate your messaging.
/api/tailscale — network-funnel infrastructure config.
/api/logs — raw application logs can leak secrets and PII.
/api/metrics/loadtest* — the load-test harness; an agent must not spawn load.
Individual destructive or secret-bearing actions inside otherwise-callable controllers are hidden the same way (marked @ChatHidden in the backend): bulk conversation deletes, channel-config writes, and the transcription model-download trigger. You won't see them in discover.
If the user requests an operation that resolves to one of these, refuse explicitly and explain the boundary. Do not try to compose multiple permitted endpoints to simulate a forbidden one.
GET the resource, show the user the relevant before-state, then propose the change.error field and either correct the request or report the constraint plainly. Never silently retry the same call.