Harmonist is a multi-agent framework for coding assistants that uses hooks to mechanically check required reviews, memory updates, and file supply-chain records before a code-changing turn can finish. It is for teams using tools such as Cursor, Claude Code, Copilot, Windsurf, or Aider to coordinate agents under enforced protocols. The catalogue entries provide agents and rules for Harmonist workflows.
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 agents/gammalabtechnologies/harmonist/tagsgit clone --depth 1 https://github.com/GammaLabTechnologies/harmonistWrote 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/agents/gammalabtechnologies/harmonist/tags)<a href="https://agentmods.dev/agents/gammalabtechnologies/harmonist/tags"><img src="https://agentmods.dev/badge/agents/gammalabtechnologies/harmonist/tags.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 | $0.00000 | $0.01970 |
| Opus 5 | $0.00000 | $0.00985 |
| Sonnet 5 | $0.00000 | $0.00394 |
| Haiku 4.5 | $0.00000 | $0.00197 |
Grade A, and why
TAGS 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 5d 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 — 216 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tag Vocabulary
Curated list of every tag an agent is allowed to declare. Source of truth:
tags.json. Linter rejects any tag not in this list.
Why a fixed vocabulary
Auto-generated tags (json, JSON, Json, json-parsing) fragment the
routing index. A curated list makes the orchestrator's job tractable:
when it looks up by_tag.security, it gets every agent that should be
there, not half of them plus a scatter of secure, sec, security-audit.
New tags aren't forbidden — they require one thing: add the tag to
tags.json in the same MR as the agent that uses it. That discipline
is the vocabulary's immune system.
Layers
Every tag lives in one of six layers. An agent typically carries 5–10
tags spread across skill, domain, tech, and concern.
| Layer | What it answers | Typical count per agent |
|---|---|---|
skill |
What the agent does functionally (backend, review, growth, qa) | 1–3 |
domain |
Subject matter (fintech, blockchain, gamedev, xr) | 1–2 |
tech |
Concrete stack / platform (react, postgres, unity, tiktok) | 0–3 |
concern |
Non-functional property (security, performance, a11y) | 0–3 |
workflow |
Output mode (implementation, audit, reporting) | 0–1 |
marker |
Pack infrastructure tags (template, bootstrap) | 0–1 |
Layer summaries
skill — what the agent does
Engineering roles: backend, frontend, fullstack, devops, sre,
infra, data-engineering, data-science, ai, ml, llm,
mobile, embedded.
Design: ui-design, ux-design, ux-research, brand-design,
visual-storytelling, whimsy, prompt-design.
Testing: qa, e2e-testing, api-testing, performance-testing,
accessibility-testing, evidence-collection, reality-check.
Protocol-bound: review, audit, scout, orchestration.
Writing: technical-writing, content-creation, book-authoring.
Marketing / growth: growth, seo, aso, aeo, social-media,
community-building, livestream-commerce, localization.
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.
- 5d ago First seen · 216 lines · 0 tokens per session scan A ca610c958864
TAGS is an agent published in the GitHub repository GammaLabTechnologies/harmonist (2,258 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,970 tokens. 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 2026-08-30.
Other agents, from other repositories
agents
Agent 模块是系统的执行引擎，基于 LangGraph StateGraph 实现单个 Agent 的完整执行生命周期管理。.
coder
Software engineer — implements features, writes tests, fixes bugs.
sec-auditor
Security audit agent — scans for vulnerabilities, checks permissions, reviews secrets.
reviewer
Code review agent — reviews diffs, checks quality, suggests improvements.
agent-capabilities
Matrix of agent capabilities and their specializations.
agent-types
Complete guide to all 54 available agent types in Claude Flow.