team-coordination

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill is entirely instructional markdown providing strategies for task delegation and project management (DAG vs. Loop).
  • [NO_CODE]: No executable scripts, configuration files, or external dependencies are included or referenced.
  • [PROMPT_INJECTION]: The instructions establish a leadership persona for the agent without attempting to bypass safety protocols or override system instructions.
  • [DATA_EXFILTRATION]: No access to sensitive file paths, environment variables, or unauthorized network endpoints was identified.
  • [INDIRECT_PROMPT_INJECTION]: The skill manages data from external sources (Workers and requesters) but includes defensive measures.
  • Ingestion points: External requester messages and Worker results processed through project-management and task-management tools (SKILL.md).
  • Boundary markers: Explicitly instructs the agent to convert raw messages into bounded tasks and avoid direct forwarding of unvalidated input.
  • Capability inventory: Uses project-management, task-management, file-sharing, organization, and communication tools (SKILL.md).
  • Sanitization: Mandates that the agent evaluate results and confirm acceptance criteria before advancing project states.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 11:25 PM
Security Audit — agent-trust-hub — team-coordination