og-image-creator
Fail
Audited by Gen Agent Trust Hub on Feb 14, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION] (HIGH): The skill directs the agent to execute local Python scripts (
analyze_codebase.pyandgenerate_og_images.py) against user-provided file paths. This grants the skill broad access to the filesystem and execution environment. Evidence: SKILL.md and README.md execution instructions. - [PROMPT_INJECTION] (HIGH): High risk of Indirect Prompt Injection (Category 8). The skill reads untrusted data from the user's project (titles, descriptions, config) to generate output. Evidence Chain: (1) Ingestion:
tailwind.config.js,page.tsx,page.astro(2) Boundary markers: Absent (3) Capability: Subprocess calls, file writes, browser rendering (4) Sanitization: Absent. - [EXTERNAL_DOWNLOADS] (LOW): Technical requirements specify downloading Chromium binaries via Playwright. While common for this tool, it remains an external code dependency. Evidence: README.md requirements section.
Recommendations
- AI detected serious security threats
Audit Metadata