swarm
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary purpose is to shard work across multiple parallel 'lanes' (agents) and then merge the results. This creates a surface where the coordinator agent processes data generated by other agents, which could contain malicious instructions.
- Ingestion points: According to the 'Merge protocol' in SKILL.md, the coordinator 'Read artifacts and changed files' from sub-agents to reconcile overlaps.
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the coordinator processes lane outputs.
- Capability inventory: The coordinator possesses the capability to integrate code changes, resolve findings, and execute 'final verification' checks (SKILL.md).
- Sanitization: There are no explicit instructions for the coordinator to sanitize, escape, or validate the content of the artifacts received from worker lanes before processing them.
Audit Metadata