manus
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates communication with the api.manus.ai domain for task creation and result retrieval, which is essential for its primary function.
- [CREDENTIALS_UNSAFE]: Uses the MANUS_API_KEY environment variable for authentication, preventing the leakage of secrets in the skill's source code.
- [COMMAND_EXECUTION]: Employs standard curl and jq commands to interact with the external API and process its responses, which is appropriate for a system-integrated skill.
- [DATA_EXFILTRATION]: While it sends user input to an external service, this behavior is restricted to the intended interaction with the Manus AI agent and does not access unauthorized local data.
Audit Metadata