← All setups
Per-agent setup
Hermes
Multi-agent orchestratorIn your stackLocal operator loop
A local loop: one model scopes a task, Claude Code builds it, audits gate the result, and every run ends with a summary you can parse.
Best forBounded, hands-off build runs — often overnight — with checks instead of blind trust.
---HERMES_SUMMARY--- status: continue | done | blocked did: - what you accomplished next: - what should happen next run audits: code: ship | needs_fixes | not_run ui: ship | needs_polish | not_run ---END---