triage
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that involves processing untrusted external inputs such as bug reports and web content. However, it includes defensive protocols to mitigate related risks.
- Ingestion points: Processed data includes 'web content and external data' accessed via tools like BrowserOS as described in SKILL.md.
- Boundary markers: The skill mandates the use of specific operational modes (BUILD, INCIDENT, EXPERIMENT) and requires the redaction of secrets and PII.
- Capability inventory: The skill utilizes an 'ExecutionProxy' for terminal commands (e.g., rtk npm test) and a 'MemoryStore' for saving persistent markdown artifacts.
- Sanitization: The instructions explicitly require treating external data as hostile and redacting sensitive information before sharing context with subagents.
Audit Metadata