deepwiki

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill references a local JavaScript file located at ./scripts/deepwiki.js for all command executions. However, this file was not provided for analysis. The security profile of the script's internal logic remains unverified.
  • [EXTERNAL_DOWNLOADS]: The skill is configured to interact with a remote API at https://mcp.deepwiki.com/mcp. This behavior is documented and consistent with the skill's purpose of fetching external documentation.
  • [PROMPT_INJECTION]: The skill provides an interface for the agent to ingest untrusted data via user-provided questions and repository paths.
  • Ingestion points: Arguments passed to the node ./scripts/deepwiki.js command (specifically <owner/repo>, "your question", and <path>).
  • Boundary markers: No delimiters or instructions to ignore embedded commands are present in the provided CLI examples.
  • Capability inventory: The skill executes local commands using Node.js and performs network requests to the DeepWiki API.
  • Sanitization: There is no evidence of input sanitization or validation within the markdown documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 05:13 AM