task-decomposer
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains no hardcoded credentials, secrets, or sensitive file paths. All templates and references are focused on structural task management.\n- [SAFE]: No unauthorized network operations, data exfiltration patterns, or external downloads from untrusted sources were identified. The skill operates within the local agent context using provided instructions and templates.\n- [SAFE]: The protocol includes a manual human-in-the-loop checkpoint (Step 3: Present the Plan), which acts as a robust mitigation against potential indirect prompt injection or unintended automated behavior.\n- [SAFE]: The skill's execution logic includes safety-conscious integration with external guardrail skills like 'checkpoint-guardian' to monitor high-risk operations and 'multi-brain' for complex decision-making.
Audit Metadata