wikipedia-research

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues were identified. The skill implements legitimate networking research functionality using established patterns for tool execution within the agent framework.
  • [COMMAND_EXECUTION]: The skill executes local Python commands to interact with its Wikipedia retrieval script. This is the standard and intended behavior for this type of skill.
  • [PROMPT_INJECTION]: An indirect prompt injection surface exists as the skill retrieves data from Wikipedia. (1) Ingestion points: Wikipedia article content retrieved via get_content, get_summary, and search_pages tools. (2) Boundary markers: No explicit delimiters or warnings are present in the provided examples. (3) Capability inventory: The agent can execute local commands via its MCP interface. (4) Sanitization: No explicit filtering of external content is performed. This surface is expected for information retrieval tools and is assessed as safe in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 12:32 AM