dispatching-parallel-agents
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions focus on task isolation, efficiency, and debugging workflows. No override, bypass, or safety filter removal patterns were detected.
- [DATA_EXFILTRATION]: No network operations, sensitive file access (such as credentials or environment files), or hardcoded secrets were found.
- [REMOTE_CODE_EXECUTION]: The skill does not download or execute scripts from remote sources. The code examples provided are instructional snippets for agent orchestration within a controlled environment.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow where agents ingest content from test files (e.g.,
src/agents/agent-tool-abort.test.ts). This presents an attack surface where malicious content inside a test file could attempt to influence agent behavior. However, this is the primary purpose of the skill, and no specific exploitable patterns or missing sanitization routines were identified that deviate from standard development practices. Boundary markers are absent in the provided example, but the scope of work is strictly constrained to specific files.
Audit Metadata