mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-11-06 05:17:15 +00:00
37 lines
1.9 KiB
YAML
37 lines
1.9 KiB
YAML
name: local-system-one-hermes
|
|
title: Local System One
|
|
repo: https://github.com/NMTZ-z/system-one-benchmark-lab
|
|
sha: 1095495019b2dff2b4231ed70ec23a3ea716e2f6
|
|
subdir: integrations/hermes/local-system-one-hermes
|
|
description: >-
|
|
Local System One / Laya Typed Decisions control plane for Hermes Agent. Sends bounded turn data to
|
|
a configurable sidecar for reversible OFF/Shadow/Canary routing decisions, with explicit Canary
|
|
acknowledgement and fail-open behavior. Disclosure — executes nothing and does not start the
|
|
sidecar. In Shadow/Canary mode, Search and Model Tier may POST the current task plus bounded recent
|
|
conversation context; Notification Shadow may POST the completed turn's final assistant response
|
|
(up to 6000 chars) plus lifecycle metadata. The default sidecar URL is loopback
|
|
127.0.0.1:8787 but is user-configurable to any host and no auth header is added. Notification is
|
|
observation-only: the plugin never sends a push/message/webhook and persists decision metadata,
|
|
not raw response text. Its only request mutations, in acknowledged Canary mode on the first
|
|
provider call of a turn, are removing web_search/web_extract tools for audited deterministic
|
|
no-Web rules and (off by default) lowering reasoning_effort for gemini-3.8-flash-tiered on audited
|
|
hard-fast rules. The sidecar (same repo, separate pip install, Apple-Silicon MLX/CoreML) is a text
|
|
classifier that downloads a pinned Hugging Face model.
|
|
maintainer: NMTZ-z
|
|
tier: community
|
|
category: automation
|
|
requires_hermes: ">=0.21.4"
|
|
version: "0.6.0"
|
|
docs_url: https://github.com/NMTZ-z/system-one-benchmark-lab/tree/1095495019b2dff2b4231ed70ec23a3ea716e2f6/integrations/hermes/local-system-one-hermes
|
|
readme: true
|
|
platforms: [macos]
|
|
capabilities:
|
|
provides_tools: []
|
|
provides_hooks:
|
|
- pre_llm_call
|
|
- post_llm_call
|
|
- on_session_end
|
|
provides_middleware:
|
|
- llm_request
|
|
requires_env: []
|