mcp-ops
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides developer documentation and utility scripts for maintaining Model Context Protocol servers. The included maintenance script (check-mcp-facts.py) probes official package registries (npm, PyPI) to ensure SDK references remain current, which is a benign administrative task.\n- [INDIRECT_PROMPT_INJECTION]: The skill documents tools and resources that process data from external sources such as web pages, databases, and files, which is an inherent attack surface for indirect prompt injection.\n
- Ingestion points: Examples in references/tool-handlers.md (e.g., fetch_webpage, query_sqlite) and references/resources-prompts.md (e.g., get_weather).\n
- Boundary markers: Not explicitly used in documentation snippets, though the skill provides instructions on implementing validation logic.\n
- Capability inventory: The skill patterns include file system access (read_file, write_file), network operations (httpx), and database execution (sqlite3).\n
- Sanitization: The documentation promotes security best practices, including directory traversal prevention using os.path.realpath and schema validation using Pydantic and Zod.
Audit Metadata