tools-page-seo-optimizer
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute several Bash and Python snippets to audit the codebase and validate SEO data. Key examples include a validation script that reads data/tools.json to verify metadata lengths and a verify function that uses curl to check site status.
- [EXTERNAL_DOWNLOADS]: Performs network requests via curl to fetch and inspect the content of the target website (e.g., sitemap.xml, tool pages) for verification purposes. These operations target user-specified or environment-defined domains.
- [INDIRECT_PROMPT_INJECTION]: The skill processes project data files, creating a potential surface for indirect injection. 1. Ingestion points: data/tools.json and data/slugs.txt (referenced in references/detailed-guide.md). 2. Boundary markers: No explicit delimiters or instructions to ignore embedded content are used. 3. Capability inventory: curl, grep, python3 execution, and build commands (e.g., npm run build). 4. Sanitization: No sanitization of the data files is performed before they are processed by the scripts.
Audit Metadata