improve-folder-architecture
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill generates HTML reports that load necessary assets from well-known content delivery networks, specifically Tailwind CSS from cdn.tailwindcss.com and Mermaid.js from jsdelivr.net.
- [COMMAND_EXECUTION]: The instructions direct the agent to use the platform's default 'open' or equivalent command to show the generated HTML report to the user, which is a standard functional requirement for visual reports.
- [PROMPT_INJECTION]: The skill processes project-specific data (folder names, READMEs, and git logs) into an HTML report, presenting an indirect prompt injection surface.
- Ingestion points: Repository root contents, manifests, and developer documentation are read in SKILL.md Section 1.
- Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the ingested metadata.
- Capability inventory: The skill creates a local file and triggers a system open command as described in SKILL.md Section 4.
- Sanitization: The Mermaid.js configuration in HTML-REPORT.md uses securityLevel: 'loose', which allows the rendering of HTML tags potentially present in malicious repository metadata.
Audit Metadata