agent-flight-recorder
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface (Category 8).
- Ingestion points: The skill captures external data via error signatures (
signal), task descriptions, and situational context inSKILL.md. - Boundary markers: Data is delimited using fenced YAML blocks with specific quoting requirements for special characters.
- Capability inventory: The skill utilizes file system tools to create directories and append log data to
.agents/flight-recorder/as defined inSKILL.md. - Sanitization: The skill provides manual instructions for the agent to anonymize content and exclude PII, tokens, or secrets, but lacks automated validation or sanitization mechanisms beyond basic YAML structure.
Audit Metadata