local-system-one-hermes.yaml 1.9 KB

123456789101112131415161718192021222324252627282930313233343536
  1. name: local-system-one-hermes
  2. title: Local System One
  3. repo: https://github.com/NMTZ-z/system-one-benchmark-lab
  4. sha: 1095495019b2dff2b4231ed70ec23a3ea716e2f6
  5. subdir: integrations/hermes/local-system-one-hermes
  6. description: >-
  7. Local System One / Laya Typed Decisions control plane for Hermes Agent. Sends bounded turn data to
  8. a configurable sidecar for reversible OFF/Shadow/Canary routing decisions, with explicit Canary
  9. acknowledgement and fail-open behavior. Disclosure — executes nothing and does not start the
  10. sidecar. In Shadow/Canary mode, Search and Model Tier may POST the current task plus bounded recent
  11. conversation context; Notification Shadow may POST the completed turn's final assistant response
  12. (up to 6000 chars) plus lifecycle metadata. The default sidecar URL is loopback
  13. 127.0.0.1:8787 but is user-configurable to any host and no auth header is added. Notification is
  14. observation-only: the plugin never sends a push/message/webhook and persists decision metadata,
  15. not raw response text. Its only request mutations, in acknowledged Canary mode on the first
  16. provider call of a turn, are removing web_search/web_extract tools for audited deterministic
  17. no-Web rules and (off by default) lowering reasoning_effort for gemini-3.8-flash-tiered on audited
  18. hard-fast rules. The sidecar (same repo, separate pip install, Apple-Silicon MLX/CoreML) is a text
  19. classifier that downloads a pinned Hugging Face model.
  20. maintainer: NMTZ-z
  21. tier: community
  22. category: automation
  23. requires_hermes: ">=0.21.4"
  24. version: "0.6.0"
  25. docs_url: https://github.com/NMTZ-z/system-one-benchmark-lab/tree/1095495019b2dff2b4231ed70ec23a3ea716e2f6/integrations/hermes/local-system-one-hermes
  26. readme: true
  27. platforms: [macos]
  28. capabilities:
  29. provides_tools: []
  30. provides_hooks:
  31. - pre_llm_call
  32. - post_llm_call
  33. - on_session_end
  34. provides_middleware:
  35. - llm_request
  36. requires_env: []