Imported from theonlygeranium/Project-Tango (.cursor/skills/cursor-discord-docs/SKILL.md). Install upstream with npx skills add theonlygeranium/Project-Tango --skill cursor-discord-docs. Copyright stays with the author.

Cursor & Discord docs retrieval

This skill ensures the agent consults the current published documentation for Cursor and Discord before generating code, rather than relying on possibly- stale model memory.

Libraries and their Context7 IDs

Procedure

1. Identify the target. Is this about Cursor internals or the Discord API? Pick the matching library ID from the table above.
2. Query with a specific question. Good: "How to configure an SSE MCP server in mcp.json." Bad: "mcp". The Context7 API ranks by relevance, so specificity improves results.
3. Read the returned snippets. Each result includes a source URL. Use those URLs as citations when explaining the code to the user.
4. Write code consistent with the retrieved docs. If the docs contradict your memory, follow the docs and note the discrepancy.
5. Stop after 3 calls per question. Reuse earlier results rather than re-querying. If you still lack what you need, fall back to your knowledge and flag the uncertainty.

What to avoid

• Do not paste secrets into the query field. Context7 queries are sent to a third-party API.
• Do not treat the docs MCP as authoritative for anything beyond Cursor and Discord. It is a context-enrichment tool, not a general oracle.
• Do not silently ignore empty results. Report them so the user can decide whether to proceed from memory.

Example flow

User: "Add a Discord bot that connects to the gateway and logs READY."

Agent:
1. Call Context7 QUERY_DOCS with libraryId=/discord/discord-api-docs
   and query="How to authenticate a bot and connect to the gateway".
2. Receive: bot token auth header, Get Gateway URL, wss connection string,
   Identify/Heartbeat event list.
3. Write the bot using those specifics, citing the source URLs.