tool-grok-search

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 external, attacker-controlled sources.
  • Ingestion points: Untrusted data enters the agent context via the grok_search tool output (X posts, news articles, web content).
  • Boundary markers: The skill definition does not specify delimiters or instructions to treat search results as untrusted data.
  • Capability inventory: The tool itself is for retrieval, but its output (the 'Answer') is intended to influence agent reasoning and downstream decisions.
  • Sanitization: No sanitization or filtering of external content is mentioned in the provided documentation.
  • External References (LOW): The documentation references local source files (src/agents/grok.ts, src/tools/grok-search.ts) that were not provided for analysis. This limits the ability to verify implementation-level security such as API key handling or input validation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 01:33 AM