tahr-test-access-control
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of instructional content for security professionals and AI agents. No executable scripts, hidden code, or malicious command patterns were identified.\n- [DATA_EXFILTRATION]: The skill implements best practices by requiring the redaction of sensitive data (credentials, tokens, cookies, PII) from analysis reports and runtime processing.\n- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface as it ingests untrusted data from API origins and traffic modeling (SKILL.md). However, the risk is mitigated by explicit safety boundaries, the requirement for fresh disposable objects for mutations, and a mandatory 'Five Proof Gates' validation process to ensure impact is proven rather than inferred from data.
Audit Metadata