docs-create
Pass
Audited by Gen Agent Trust Hub on Feb 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a strict 'Human Review Check' between analysis and generation phases, preventing the agent from creating or modifying files without explicit user consent.
- [SAFE]: Permissions are minimized using scoped tool configurations. The Bash tool is restricted to 'git add' and 'git commit' with wildcards, and the Write tool is constrained to documentation and output directories ($JAAN_DOCS_DIR and $JAAN_OUTPUTS_DIR).
- [SAFE]: The workflow includes a validation step that checks generated content against a local STYLE.md file to ensure adherence to formatting rules and length limits before the file is written.
- [SAFE]: The skill presents an indirect prompt injection surface (Category 8) by processing user arguments and local files. 1. Ingestion points: $ARGUMENTS and files in $JAAN_DOCS_DIR. 2. Boundary markers: none defined. 3. Capability inventory: restricted file writing, git commits, and config editing. 4. Sanitization: structural validation via STYLE.md. This is documented as a standard functional surface.
- [SAFE]: All operations are performed locally using project-specific templates and protocols located within the plugin or project structure. No remote downloads or network operations are detected.
Audit Metadata