agent-orchestration

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to launch and interact with various CLI-based AI tools and session managers (e.g., wsh server, claude, aider, git worktree).
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes and reacts to instructions or output from external, untrusted AI agents.
  • Ingestion points: The agent is directed to read terminal output and scrollback from child agent sessions (SKILL.md).
  • Boundary markers: The skill suggests using pattern matching for specific terminal states (e.g., "Approval prompts" or "Waiting for input") but does not provide robust delimiters to separate data from instructions.
  • Capability inventory: The agent has access to wsh tools which can send arbitrary input to terminal sessions and manage processes.
  • Sanitization: No explicit sanitization or validation of the output received from sub-agents is performed before the primary agent makes decisions based on that output.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 09:56 AM
Security Audit — agent-trust-hub — agent-orchestration