improve-codebase-architecture
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses platform-specific commands (xdg-open, open, or start) to display the generated architectural report to the user in their default browser.
- [EXTERNAL_DOWNLOADS]: The generated HTML report fetches style and diagramming resources from well-known services, specifically Tailwind CSS (tailwindcss.com) and Mermaid.js via jsDelivr (jsdelivr.net).
- [PROMPT_INJECTION]: The skill reads local project data and git history to generate a report, creating a potential indirect prompt injection surface where malicious codebase content could influence the generated output.
- Ingestion points: Reads codebase files, git log output, CONTEXT.md, and Architecture Decision Records (ADRs) from docs/adr/.
- Boundary markers: The instructions do not define clear boundaries or provide explicit guidance to the agent for sanitizing external content before interpolation into the HTML report.
- Capability inventory: The skill has the ability to write to the OS temporary directory and execute shell commands to open generated files.
- Sanitization: No sanitization is performed. The report template explicitly configures Mermaid.js with securityLevel: "loose", which permits the execution of script tags or HTML elements contained within the processed codebase data.
Audit Metadata