jules
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local shell commands (
git,sed,awk,grep) to extract repository context and construct task descriptions. These operations are standard workflow helpers for development tools. - [EXTERNAL_DOWNLOADS]: The skill provides instructions to download the official Google Jules CLI via
npm install -g @google/jules. These installations utilize standard registries and target a well-known service ecosystem, making them safe. - [DATA_EXFILTRATION]: Network requests (
curl) are exclusively targeted at official Google API endpoints (jules.googleapis.com), passing authenticating parameters securely via standard headers (x-goog-api-key) with no signs of credential harvesting or unsafe external redirection. - [INDIRECT_PROMPT_INJECTION]: The skill provides examples of injecting dynamic context fields like
PR_INFOusing the GitHub CLI (gh pr view). The surface handles external content via standard API structures and wraps queries inside bounded templates.
Audit Metadata