ae-system

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [SAFE]: The skill instructions define administrative workflows for managing the Agent environment and explicitly incorporate security guardrails such as ID verification and dry-run requirements.
  • [COMMAND_EXECUTION]: The skill relies on the ae-cli system for all operations. Instructions emphasize adhering to authorization boundaries and correctly handling permission denials by verifying the user's root or agent_admin role.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied rosters and external API data from Feishu, it mitigates potential injection risks through mandatory identity discovery, validation steps, and human-in-the-loop confirmation for every write operation.
  • [DYNAMIC_EXECUTION]: The skill facilitates the installation of npm packages and the upload of custom tools into sandboxes. Security is maintained through default script blocking (--ignore-scripts), exact version pinning, path normalization, and manifest validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 02:32 AM
Security Audit — agent-trust-hub — ae-system