orchestrating-agent-delegation
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a documentation-based resource providing guidelines and templates for multi-agent orchestration without any executable scripts or network-active commands.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that interpolates external data such as repository code and specifications into agent context. It includes several mitigation strategies for the associated risks:
- Ingestion points: The prompt template in
SKILL.mdidentifiesSPECandREPOsections as points where external file content and repository state enter the agent context. - Boundary markers: The skill explicitly recommends limiting ingestion to specific line ranges and isolated files to prevent context contamination.
- Capability inventory: Subagents are expected to have the capability to modify the local file system and execute shell commands (e.g., tests).
- Sanitization: The workflow enforces sanitization through a 'trust-but-verify' discipline (manual diff review) and the use of 'refute-first' adversarial review panels to validate subagent findings before they are accepted by the orchestrator.
Audit Metadata