pentest-outbound-interaction-oob-detection
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external network interactions to validate security findings, creating a surface for indirect prompt injection.
- Ingestion points: The skill reads callback data from
interactions.jsonl(generated by theinteractsh-clientlistener) to correlate tokens and confirm findings inSKILL.md. - Boundary markers: The instructions do not specify the use of delimiters or boundary markers when the agent processes or reports the content of the observed interactions.
- Capability inventory: The skill uses shell execution to start the background listener and file system operations to store and read interaction logs in the
/tmpdirectory. - Sanitization: There is no explicit requirement to sanitize or filter the content of the callback data (e.g., HTTP headers or bodies) before it is incorporated into the agent's reasoning or output table.
Audit Metadata