improve-codebase-architecture

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses system commands to open generated files via open, xdg-open, or start and interacts with the local environment using git and the GitHub CLI (gh) for repository exploration and issue creation.
  • [EXTERNAL_DOWNLOADS]: Generated HTML reports load JavaScript and CSS from external CDNs, specifically cdn.tailwindcss.com and cdn.jsdelivr.net.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the analyzed codebase, including ADRs, glossaries, and source files, to generate reports and issues. Ingestion points include file reads across the repository; capability inventory includes local file writes (temp reports) and GitHub issue creation; no explicit boundary markers or sanitization steps are defined for handling the external codebase content.
  • [DYNAMIC_EXECUTION]: The Mermaid diagram engine in the generated HTML reports is configured with securityLevel: "loose". This setting allows the rendering of HTML tags in labels, which could lead to browser-based script execution if module names or descriptions in the codebase contain malicious payloads.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 06:30 PM
Security Audit — agent-trust-hub — improve-codebase-architecture