ns-harness-agents-md
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed to generate project entry documents (
AGENTS.mdandCLAUDE.md) based on a read-only scan of the repository's layout, skills, and documentation. - [SAFE]: It employs a template-driven approach with strict output constraints, such as line limits and specific mandatory sections, which reduces the risk of malicious instruction generation.
- [SAFE]: The skill uses the vendor's own package (
@nextstage-brasil/harness) for specific operations, which is an expected behavior for a skill from this author. - [SAFE]: Explicit safety constraints are included in the generated
AGENTS.mdfile, such as forbidding the mutation of GitLab state or restarting Docker containers without user confirmation.
Audit Metadata