bug-hunt-swarm
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the workspace which could contain instructions designed to manipulate the agent's behavior.
- Ingestion points: In SKILL.md (Step 1), the agent collects symptoms, logs, stack traces, and relevant code files from the local environment.
- Boundary markers: The instructions do not explicitly specify the use of delimiters or 'ignore previous instructions' markers when including untrusted evidence in prompts for the sub-agents.
- Capability inventory: The skill utilizes search and version control tools (
rg,git diff,git log,git show) and orchestrates four parallel sub-agents. While the sub-agents are restricted to read-only roles, the main agent synthesizes and acts on their findings. - Sanitization: There is no evidence of sanitization or validation of the ingested content before it is processed by the AI models.
Audit Metadata