qws-interview
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input during the interview process and from imported interview records. This content is used to generate structured knowledge drafts and complete interview logs stored as Markdown files. The risk of malicious instructions in user data influencing agent behavior is mitigated by the 'Contract and Writer' protocol, which uses independent sub-agents for content analysis and synthesis to provide a security boundary.- [DYNAMIC_EXECUTION]: The skill utilizes a local Python script (
scripts/verify_theory_overlay.py) to perform automated integrity checks on generated documents. This script verifies that the 'theory overlay' (external references) has not modified the original frozen text. The script uses standard libraries and safe file-handling practices.
Audit Metadata