parallel-dispatch
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines architectural guidelines for parallel agent execution, focusing on MECE task decomposition and ownership invariants. No malicious command execution, data exfiltration, or persistence mechanisms were found.
- [INDIRECT_PROMPT_INJECTION]: The skill describes an 'Agent Spawn Protocol' which involves interpolating workspace paths and agent types into system prompts for sub-agents. This presents a potential surface for indirect injection, though it is standard for orchestration logic.
- Ingestion points: Work deliverables, dependency mappings, and agent type identifiers processed in
SKILL.md. - Boundary markers: None explicitly defined for the interpolation of variables into the sub-agent system prompts.
- Capability inventory: The skill manages the spawning of sub-agents and enforces file ownership through glob-based write scopes.
- Sanitization: No explicit path sanitization or validation is mentioned for the agent role file references.
Audit Metadata