2.6k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
3 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill comind- _meta.json264 B
- SKILL.md37.2 KB
Overview
This skill documents the standardized runtime behavior for an AI member on the CoMind collaboration platform. It defines task execution, Markdown sync, conversation actions, status panel updates, and MCP API usage as the reliable fallback. Use this skill when the agent receives task pushes, chat requests, scheduled jobs, or inspection directives from CoMind.
How this skill works
The skill describes three interaction channels: Conversation Actions (fast, WebSocket-dependent), MCP API (reliable HTTP fallback with explicit errors and audit), and Markdown Front Matter sync (bulk convenience). It provides decision rules, verification scripts, and retry patterns so the agent first attempts efficient channels and always verifies critical changes through MCP API. Built-in shell templates show how to call MCP methods, verify results, and handle rate limits or transient failures.
When to use it
- Use Conversation Actions for quick in-chat updates that can be retried or verified later.
- Use MCP API whenever a 100% confirmation of result, audit trail, or query is required.
- Use Markdown Front Matter sync for bulk creating or updating multiple tasks or deliveries.
- Combine Markdown or Actions with MCP verification for reliable end-to-end outcomes.
- Prefer MCP API for schedule management, queries, or any operation Actions do not support.
Best practices
- Always verify document syncs or Actions that affect critical state by calling MCP get/list APIs.
- Handle transient errors and rate limits with exponential backoff and limited retries.
- Use idempotency keys for actions to avoid duplicate side effects during retries.
- Keep Front Matter format exact and validate locally before uploading to reduce silent parsing failures.
- Log all MCP responses and errors for auditability and troubleshooting.
Example use cases
- Create multiple tasks via a Markdown document with type: comind:tasks, then call list_my_tasks to confirm counts.
- Update a task status from chat using an embedded actions JSON, then call get_task to verify the change.
- Submit a document with delivery_status: pending in Front Matter, then call list_my_deliveries and get_delivery to confirm the record.
- Run periodic inspections that create or update tasks via Markdown sync and validate results with MCP API.
- Manage schedules or run queries exclusively through MCP API since Actions cannot perform scheduling or queries.
FAQ
Always trust MCP API for final verification; it provides explicit errors, audit logs, and supports full query and write operations.
What if Markdown Front Matter parsing fails silently?
Treat Front Matter sync as a convenience only. Immediately call MCP list/get APIs to confirm created records and handle any discrepancies.