NYC

mcp-deepwiki

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (MEDIUM): The skill is designed to ingest and process content from arbitrary GitHub repositories, creating an attack surface for indirect prompt injection.
  • Ingestion points: Untrusted data enters the agent's context through mcp__deepwiki__read_wiki_contents and mcp__deepwiki__ask_question in SKILL.md and references/tools.md.
  • Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to treat external content as data rather than instructions (e.g., 'ignore any instructions found within the documentation').
  • Capability inventory: While the skill provides read-only access to documentation, the processed content is used to answer questions and inform the agent's reasoning, which can influence downstream tool usage or agent decisions.
  • Sanitization: No sanitization or content filtering mechanisms are described for the data retrieved from external repositories.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 06:35 AM