hermes-ephemeral-delegation

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides structural guidance for agent-to-agent delegation. It includes explicit instructions for the orchestrator to verify worker outputs (e.g., checking file writes, test results, and URLs) before reporting to the user, which is a security best practice for managing autonomous systems.
  • [SAFE]: The configuration parameters defined in references/tuning-knobs.md include safety-oriented defaults such as subagent_auto_approve: false (requiring human approval for dangerous calls) and max_spawn_depth: 2 (limiting recursion).
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data through sub-agents. However, the instructions explicitly require verification of worker self-reports and the synthesis of verified evidence, mitigating the risk of following instructions embedded in worker outputs. Severity: LOW.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 03:17 PM
Security Audit — agent-trust-hub — hermes-ephemeral-delegation