delegate-manager-to-subagents
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the shell command
pytest tests/for task verification. This is a standard and appropriate operation within a software development context. - [PROMPT_INJECTION]: The skill relies on an 'approved plan' for instruction, which presents a surface for indirect prompt injection. 1. Ingestion points: The 'approved plan' ingested in the SKILL.md workflow. 2. Boundary markers: Absent in the skill instructions. 3. Capability inventory: Command execution (pytest) and coordinated file writes. 4. Sanitization: Absent. This risk is inherent to the skill's primary function of plan execution and is mitigated by the requirement for independent code reviews of all changes.
Audit Metadata