spec-quality

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill is purely informational, consisting of markdown instructions and evaluation rubrics for reviewing specifications.
  • [NO_CODE]: No scripts, shell commands, binaries, or configuration files are present in the skill package.
  • [DATA_EXFILTRATION]: No network access patterns or commands (e.g., curl, wget) are used, and no sensitive file paths are accessed.
  • [PROMPT_INJECTION]: The instructional content defines evaluation criteria without attempting to bypass safety protocols or override agent behavior instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 10:40 AM
Security Audit — agent-trust-hub — spec-quality