improve-codebase-architecture
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the codebase and renders it into a generated HTML report.\n
- Ingestion points: Codebase files and
git loghistory are read during the 'Explore' step inSKILL.md.\n - Boundary markers: The skill does not provide specific delimiters or instructions to prevent content from the codebase from being interpreted as active HTML/JS within the report.\n
- Capability inventory: The skill uses system commands (
xdg-open,open, orstart) to automatically trigger the browser to open the generated report file as described inSKILL.md.\n - Sanitization: The report template in
HTML-REPORT.mdexplicitly configures the Mermaid diagramming tool withsecurityLevel: "loose", which allows HTML tags to be rendered. This creates a potential cross-site scripting (XSS) surface if filenames or code comments containing scripts are incorporated into the report diagrams or text.\n- [COMMAND_EXECUTION]: The skill utilizes standard system commands to perform analysis and display results.\n - Evidence: The skill executes
git log --onelineto identify hotspots and usesxdg-open,open, orstartto present the generated architectural review to the user inSKILL.md.
Audit Metadata