auth-debug

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [Indirect Prompt Injection] (LOW): Ingestion points: specs/specs.md, specs/security-spec.md, and request headers/cookies. Boundary markers: Absent. Capability inventory: None. Sanitization: Absent. The skill processes untrusted external data, but because it has no executable functionality or network access, the risk is minimal.
  • [Data Exposure] (SAFE): The skill references security-related specification files. Accessing these files is necessary for the skill's primary purpose and no patterns of exfiltration or hardcoded credentials were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 01:32 AM