discord
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill defines a 'read' action that allows the agent to ingest untrusted message content from Discord channels, creating a surface for indirect prompt injection attacks.
- Ingestion points: Discord message retrieval via the 'read' action in SKILL.md.
- Boundary markers: The skill does not define delimiters or specific instructions to isolate message content from the agent's command context.
- Capability inventory: The skill allows significant actions including message sending, deletion, polling, and searching as documented in SKILL.md.
- Sanitization: No sanitization or validation logic is specified for the ingested content.
- [SAFE]: The skill follows security best practices for secret management by requiring a 'channels.discord.token' from the platform configuration instead of using hardcoded credentials.
- [SAFE]: The instructions explicitly mention respect for permission gating (channels.discord.actions.*), which mitigates the risk of unauthorized use of restricted Discord features like moderation or role management.
Audit Metadata