discord
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill demonstrates the use of the
file://URI scheme within themediaparameter of thesendaction (e.g.,file:///tmp/example.png). This capability could be exploited to exfiltrate sensitive local files from the host environment if the agent is manipulated into specifying unauthorized file paths. - [PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by ingesting untrusted content from external sources through
readandsearchactions. - Ingestion points: Discord message content retrieved via
action: "read"andaction: "search"inSKILL.md. - Boundary markers: None identified; instructions do not provide delimiters or warnings for the agent to ignore instructions contained within retrieved messages.
- Capability inventory: The skill allows reading from the network (Discord), writing to the network (Discord messages/reactions), and accessing the local file system (via
mediafield). - Sanitization: No sanitization, validation, or filtering of the incoming message content is defined.
Audit Metadata