cr-review

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a bash helper function (gbapi) to execute curl and jq commands for interacting with the GitBook REST API.
  • [EXTERNAL_DOWNLOADS]: The skill makes network requests to api.gitbook.com, which is the official API for GitBook, a well-known service.
  • [PROMPT_INJECTION]: The skill processes external data from GitBook change requests and comments, creating an indirect prompt injection surface. Evidence: 1. Ingestion points: GitBook API responses in SKILL.md. 2. Boundary markers: Explicit instructions in SKILL.md to treat content as data. 3. Capability inventory: Subprocess calls to curl and jq in SKILL.md. 4. Sanitization: Instructions to surface content to the user instead of acting on it.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 04:29 PM
Security Audit — agent-trust-hub — cr-review