Installable skillMIT

Discord Bot API Skill (Advanced)

OpenClaw skill for Discord Bot API workflows, covering interactions, commands, messages, and operations using direct HTTPS requests.

BY kbarbel640-delUPDATED Oct 27, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.

instructions.md

---
name: discord-hub
description: OpenClaw skill for Discord Bot API workflows, covering interactions, commands, messages, and operations using direct HTTPS requests.
---

# Discord Bot API Skill (Advanced)

## Purpose
Provide a production-oriented guide for building Discord bot workflows via the REST API and Interactions, focusing on professional command UX, safe operations, and direct HTTPS usage (no SDKs).

## Best fit
- You want command-first bot behavior and clear interaction flows.
- You prefer direct HTTP requests without a library dependency.
- You need a structured map of Discord API surfaces.

## Not a fit
- You need a full SDK or gateway client implementation.
- You plan to stream large media uploads directly.

## Quick orientation
- Read `references/discord-api-overview.md` for base URL, versioning, and object map.
- Read `references/discord-auth-and-tokens.md` for token types and security boundaries.
- Read `references/discord-interactions.md` for interaction lifecycle and response patterns.
- Read `references/discord-app-commands.md` for slash, user, and message commands.
- Read `references/discord-messages-components.md` for messages, embeds, and components.
- Read `references/discord-gateway-webhooks.md` for gateway vs webhook tradeoffs.
- Read `references/discord-rate-limits.md` for throttling and header-based handling.
- Read `references/discord-request-templates.md` for HTTP payload templates.
- Read `references/discord-feature-map.md` for the full surface checklist.

## Required inputs
- Bot token and application ID.
- Interaction endpoint public key (if using interaction webhooks).
- Command list and UX tone.
- Allowed intents and event scope.

## Expected output
- A clear bot workflow plan, command design, and operational checklist.

## Operational notes
- Prefer interactions and slash commands over prefix parsing.
- Always validate incoming interaction signatures.
- Keep payloads small and respond quickly to interactions.

## Security notes
- Never log tokens or secrets.
- Use least-privilege permissions and scopes.
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.