deepwiki

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions for the user to configure an external Model Context Protocol (MCP) server located at https://mcp.devin.ai/mcp. This is the official endpoint for Devin's documentation service (Cognition AI), a well-known technology provider for AI agent tooling.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process data from external repositories and user questions.
  • Ingestion points: Data retrieved from GitHub repositories (architecture documentation, design decisions, code intent) via the DeepWiki indexing service, as well as natural language questions passed to the mcp__devin__ask_question tool.
  • Boundary markers: The skill does not define specific boundary markers or escaping mechanisms to isolate the untrusted repository content or the user's question from the tool's execution context.
  • Capability inventory: The skill utilizes file system tools (Glob, Grep, Read) and interacts with an external AI documentation service via MCP.
  • Sanitization: There is no evidence of input validation, filtering, or sanitization of the content fetched from remote repositories before it is processed by the AI.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 03:25 PM