xai

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill incorporates a search tool (scripts/search-x.js) that retrieves and displays content from X (Twitter). This creates a surface for indirect prompt injection where malicious instructions embedded in tweets could attempt to influence the agent's behavior.
  • Ingestion points: Untrusted data enters the agent context via the search results returned by the xAI Responses API in scripts/search-x.js.
  • Boundary markers: The script does not utilize specific delimiters or provide warnings to the LLM when displaying external content.
  • Capability inventory: The skill is capable of performing network requests to api.x.ai using the https module.
  • Sanitization: Content retrieved from the API is logged directly to the console without filtering or sanitization.
  • [DATA_EXPOSURE]: The scripts (chat.js, models.js, search-x.js) read the agent's local configuration file at ~/.clawdbot/clawdbot.json to retrieve the xai API key. This is a standard configuration mechanism for skills designed for the OpenClaw/Clawdbot framework and does not represent unauthorized data access.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 07:55 AM