grok-api
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Retrieves live documentation and model metadata from official xAI domains (docs.x.ai, x.ai) and instructs the installation of the 'openai' and 'zod' packages from standard registries.
- [COMMAND_EXECUTION]: Utilizes the Bash tool and provides instructions for dependency management via pnpm. Includes code patterns that interact with the local filesystem using the Node.js 'fs' module to load context into the model.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external content from user-specified files and external URLs into the model's prompt context.
- Ingestion points: Phase 1 (reading code files specified in arguments) and Phase 2 Pattern 5 (loading document corpora into context).
- Boundary markers: Employs XML-style tags in the documentation injection pattern to separate file content from system instructions.
- Capability inventory: The agent has access to powerful tools including Bash, Write, Edit, and WebFetch, which could be abused if malicious instructions are successfully injected via processed data.
- Sanitization: No explicit logic is provided to sanitize or validate the content of the ingested files or web pages against embedded instructions.
Audit Metadata