Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/srikanthmx/hermes-cli-orchestrator/agents-mdgit clone --depth 1 https://github.com/srikanthmx/hermes-cli-orchestratorWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/srikanthmx/hermes-cli-orchestrator/agents-md)<a href="https://agentmods.dev/instructions/srikanthmx/hermes-cli-orchestrator/agents-md"><img src="https://agentmods.dev/badge/instructions/srikanthmx/hermes-cli-orchestrator/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.04327 | $0.04327 |
| Opus 5 | $0.02164 | $0.02164 |
| Sonnet 5 | $0.00865 | $0.00865 |
| Haiku 4.5 | $0.00433 | $0.00433 |
Grade A, and why
hermes-cli-orchestrator AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 181 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — Hermes CLI Governor (cli-orchestrator)
Read this before doing anything. Do not re-derive the project's purpose — it's here.
0. North Star (the whole point)
This plugin turns Hermes into a governance layer over many cheap/free brains and CLIs so the agent runs at ~$0 and never hits a wall.
The central idea: there is no single "best brain." The pool is the brain.
- Consistency and learning in Hermes are model-independent — memory, skills, session/journey persist no matter which backend answered a given call. So the brain can change under the hood on every call and it's still the same agent with the same memory and behavior.
- Therefore the goal is not "pick Codex or Ollama." It's: stack as many independent quota buckets as possible and rotate across them, auto-skipping exhausted ones, so the user never stalls and never notices the switch.
- Each of these is a separate quota bucket, and they add up:
- Subscription providers: Codex (ChatGPT), Copilot — premium but harshly capped.
- Free OAuth: Qwen-OAuth (~2000/day), Nous.
- Free API: Gemini API, OpenRouter
:free, HuggingFace, Z.ai — each its own cap. - Trial credits: NVIDIA, Novita, GMI.
- CLI workers via
cli_delegate: agy, opencode, codex-exec — more separate quota. - Multiple keys/accounts per provider — multiplies each bucket.
- Ollama — optional bonus for users who have it. NEVER the assumed floor (most users won't set it up).
The plugin "boasts of a consistent brain and a governance layer that hides the difference between switched models." Every design decision must serve that. If a suggestion relies on one model, or assumes Ollama, it's wrong.
1. Core mental model — do NOT confuse these
| Concept | What it is | Where it lives |
|---|---|---|
| Model / provider (the "brain") | an LLM endpoint Hermes calls to reason | config.yaml model + fallback_providers; needs OAuth or an API key |
| CLI (a "worker") | a subprocess the governor delegates tasks to (cli_delegate / /cli-delegate) |
detected on PATH; NOT an LLM endpoint |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 181 lines · 4,327 tokens per session scan A f0a623150e9c
hermes-cli-orchestrator AGENTS.md is an instructions file published in the GitHub repository srikanthmx/hermes-cli-orchestrator (10 stars, last pushed 1mo ago), licensed MIT. It adds 4,327 tokens to every session, about $0.0216 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
1flowbase AGENTS.md
AGENTS.md instructions for taichuy/1flowbase, covering 记忆, 用户偏好, 本项目相关skill在, 本项目 skills and 质量控制.
untether GEMINI.md
Gemini CLI instructions for littlebearapps/untether, covering untether — gemini instructions, stack & conventions, architecture, key rules and testing.
claude-tui CLAUDE.md
Claude Code instructions for slima4/claude-tui, covering claude.md, project overview, top-level scripts, tools and claude-code-statusline.
claude-tui AGENTS.md
AGENTS.md instructions for slima4/claude-tui, covering agents.md, quick commands, testing, gotchas (hard-won knowledge) and local development.
portable-agent-skills AGENTS.md
Instructions for ch1109/portable-agent-skills: Read CONTRIBUTING.md before changing a Skill.
symphony CLAUDE.md
Instructions for lrgendie/symphony, covering claude.md — symphony proje anayasası, proje nedir?, her oturumda ilk iş, belge haritası and dokunulmaz kurallar.