discord
Pass
Audited by Gen Agent Trust Hub on Apr 6, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates reading messages from Discord (action: read), which introduces a surface for indirect prompt injection. Malicious instructions embedded in Discord messages could potentially influence the agent's behavior. * Ingestion points: Message content retrieved via the message tool in SKILL.md. * Boundary markers: Absent. The instructions do not specify delimiters or warnings for the agent to disregard instructions within retrieved messages. * Capability inventory: The skill can send messages, create threads, and manage Discord state via the message tool. * Sanitization: Absent. No mention of filtering or escaping content from external sources.
- [DATA_EXFILTRATION]: The skill provides functionality to send media from the local filesystem (media: file:///tmp/example.png). While a standard feature for messaging, this allows the agent to read local files and send them to an external network service, presenting a risk of data exposure if misused.
Audit Metadata