improve-codebase-architecture
Pass
Audited by Gen Agent Trust Hub on Jul 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Uses platform-specific commands (
xdg-open,open, orstart) to automatically display the generated HTML report in the user's default browser.- [EXTERNAL_DOWNLOADS]: Fetches styling and diagramming resources from trusted CDNs (Tailwind CSS and jsDelivr) to render the architectural report.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from the user's codebase and documentation to generate reports. - Ingestion points: Project codebase (via Explore subagent), CONTEXT.md, and ADR files.
- Boundary markers: Absent; there are no instructions to delimit codebase content from the agent's generation logic.
- Capability inventory: File system write access ($TMPDIR) and shell command execution (open).
- Sanitization: Absent; the instructions do not require escaping codebase content. Furthermore, the report template in HTML-REPORT.md initializes Mermaid with securityLevel: 'loose', which allows HTML tags in labels and could lead to XSS if malicious code snippets are embedded in diagrams.
Audit Metadata