summarize-interview
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted transcripts without boundary markers. Ingestion points: User-provided interview transcripts and attached files. Boundary markers: None present in the instructions. Capability inventory: Workspace file write (markdown). Sanitization: No validation or sanitization is performed on processed transcripts.
- [EXTERNAL_DOWNLOADS]: The skill links to educational content on the productcompass.pm domain. These links serve as static documentation provided by the author and do not involve automated execution of external code.
Audit Metadata