mcp-documentation-server

Warn

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installation instructions recommend npx -y @andrea9293/mcp-documentation-server, which downloads code from an unverified npm registry account not associated with the known vendor.
  • [REMOTE_CODE_EXECUTION]: The npx command executes the fetched remote package code in the user's local environment.
  • [INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: The process_uploads tool reads and indexes markdown, text, and PDF files from the ~/.mcp-documentation-server/uploads directory. 2. Boundary markers: The skill does not specify the use of clear delimiters or instructions to ignore embedded prompts within the documentation content. 3. Capability inventory: Several search tools (search_all_documents, search_documents_with_ai) inject retrieved document fragments into the agent's context. 4. Sanitization: No sanitization or verification logic is documented to detect or remove malicious instructions from ingested documentation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 16, 2026, 10:44 PM
Security Audit — agent-trust-hub — mcp-documentation-server