generate-mintlify-reference
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. It reads documentation content from external sources, specifically CLI help docstrings and MCP tool schemas, and writes this content to MDX files for the documentation site. If these source strings contain malicious instructions, they could be persisted into the generated documentation. Evidence: Ingestion points in 'scripts/gen_cli_reference.py' (lines 37, 51) and 'scripts/gen_mcp_reference.py' (lines 92-94). No boundary markers are used to delimit external content, and sanitization is limited to basic escaping for markdown tables.
- [COMMAND_EXECUTION]: The skill facilitates the creation and execution of local Python scripts to introspect project code. It also provides a CI workflow snippet that performs environment-modifying commands such as 'pip install -e .' and 'make docs-reference', which entails executing code from the local repository. Evidence: Template scripts in 'scripts/' and the 'Docs Reference Freshness' CI job in 'references/reference-patterns.md'.
Audit Metadata