swarm
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the aggregation of output from multiple sub-agents ('workers'), which creates a potential surface for indirect prompt injection if those workers process untrusted data.
- Ingestion points: Results are read from worker output paths in the 'Phase C: Aggregate' section.
- Boundary markers: The instructions specify a structured reporting format (PASS, ISSUES, BLOCKED) and a compact result table to limit the influence of raw sub-agent content.
- Capability inventory: The skill utilizes the 'spawn_agent' tool to launch workers and reads files to aggregate data.
- Sanitization: Instructions explicitly warn 'Do not paste raw worker dumps,' which serves as a mitigation by encouraging summarized reporting over direct interpolation of worker-generated content.
Audit Metadata