export-ml-site
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the
python -m skore_skillscommand-line utility to manage the lifecycle of the documentation site, including checking status, configuring policies, and performing builds. This tool is part of the vendor's provided environment. - [EXTERNAL_DOWNLOADS]: The skill requests the installation of the
mkdocs-materialPython package through theadd-python-packagetool. This is a well-known and standard documentation theme. - [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by aggregating and rendering content from the workspace into an HTML report.
- Ingestion points: The skill reads Markdown files (
references/human_facing_prose.md), converted notebook HTML files (<stem>.nb.html), and various results (scratch/results/<stem>/) produced by previous workflow stages. - Boundary markers: There are no explicit boundary markers or warnings provided to the agent regarding the content of these processed files.
- Capability inventory: The skill uses
python -m skore_skills site buildto write files to the workspace (specifically thereport.htmland the staged documentation directory). - Sanitization: No explicit sanitization or filtering is applied to the input content before it is processed by the MkDocs build engine.
Audit Metadata