doc-sectest

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill emphasizes security best practices by including mandatory safety rules, such as requiring security tests to be run in isolated environments and prohibiting tests against production systems.\n- [COMMAND_EXECUTION]: The skill utilizes local Python and Bash scripts for artifact validation (e.g., validate_sectest.py, validate_all_tspec.sh). These scripts are internal to the project's documentation framework and do not involve remote code execution or untrusted external sources.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it is designed to ingest and process documentation artifacts from the local filesystem.\n
  • Ingestion points: The skill reads test specifications from the docs/10_TSPEC/SECTEST/ directory.\n
  • Boundary markers: No specific boundary markers or delimiters are instructed for the agent to use when reading these files.\n
  • Capability inventory: The skill has the capability to execute local scripts (python, bash) based on the files it processes.\n
  • Sanitization: The skill relies on external validation scripts to enforce schema compliance, though it does not explicitly mention sanitizing the content for embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 04:25 AM
Security Audit — agent-trust-hub — doc-sectest