メニュー
Work on a specific task assigned to you in the agent swarm
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
Use swarm scripts for bulk SDK calls, repetitive fan-out, and context-efficient data processing.
Fetch a file attached to your current task in ONE call. Use whenever a task carries an attachment (an image, PDF, or other file the requester uploaded) and you need its bytes on disk — the dispatch prompt lists attachments with a ready-to-run curl command, but if you're improvising (resumed session, follow-up task, or the recipe scrolled out of context) use this skill instead of reaching for the `agent-fs` CLI directly.
Guide for running local E2E tests with API server, Docker lead/worker containers, task creation, log verification, UI dashboard, and cleanup
Canonical AgentMail send-message API reference for swarm agents. Pins the base URL, required field names, text-only rendering workaround, BCC policy, and ready-to-copy curl / swarm-script examples so agents do not rediscover the API surface at runtime.
How to interact with Kapso WhatsApp from the swarm — read inbound webhook payloads (text AND media), fetch message history, send free-form messages within the 24h session window (and template messages outside it), mark-as-read, show the typing indicator, send reactions, download media, verify webhook signatures, and resolve contacts to swarm users. Canonical reference for ANY Kapso interaction beyond the thin `send-whatsapp-message` / `reply-whatsapp-message` MCP tools — for templates, media, reactions, typing, mark-as-read, signature verify, contact resolution, conversation history, drop to the REST recipes here. Use whenever a task references a WhatsApp message routed through Kapso, or when a workflow needs to reply on WhatsApp.
Per-app playbook for driving Gmail through Composio (toolkit slug `gmail`). Verified GMAIL_* tool slugs and argument shapes for reading, searching, sending, drafts, labels, and threads. Use alongside the `composio` hub skill whenever a task reads or sends Gmail for a connected user. Covers the metadata-first reads, the GMAIL_SEND_EMAIL HTML flag, and reply-to-thread.
| name | work-on-task |
| description | Work on a specific task assigned to you in the agent swarm |
If no taskId is provided, call poll-task to get a new task.
Get task details: Call get-task-details with the taskId.
Recall relevant memories: Use memory-search with the task description before starting any work. Past learnings, solutions, and gotchas are indexed here.
Choose your approach based on the task type:
Work on it, calling store-progress at each meaningful milestone (not just start and end — the lead monitors this).
Complete the task — see Completion below.
Call store-progress with:
status: "completed" + output: "<what you did and the result>". Output should be specific enough for the lead to assess without re-reading your work.status: "failed" + failureReason: "<what went wrong and what you tried>".Then reply "DONE" to end the session.
If interrupted by the user, adapt to their instructions. When resuming, call /skill:work-on-task <taskId> again to pick up where you left off.
If you're stuck after genuine effort (not just first failure), use /skill:swarm-chat to ask the lead for help or clarification. Don't spin — escalate.