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/notnull92/hera-agent-unity/claude-mdgit clone --depth 1 https://github.com/NotNull92/hera-agent-unityWrote 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/notnull92/hera-agent-unity/claude-md)<a href="https://agentmods.dev/instructions/notnull92/hera-agent-unity/claude-md"><img src="https://agentmods.dev/badge/instructions/notnull92/hera-agent-unity/claude-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 | $0.11102 | $0.11102 |
| Opus 5 | $0.05551 | $0.05551 |
| Sonnet 5 | $0.02220 | $0.02220 |
| Haiku 4.5 | $0.01110 | $0.01110 |
Grade A, and why
hera-agent-unity CLAUDE.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 3d 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 — 542 lines — stays where its author put it; the contents beside it link to each section on GitHub.
hera-agent-unity
See
AGENTS.mdfor the canonical cross-tool agent rules file (read natively by Codex, Claude Code, Cursor, Copilot, and 30+ tools).
CLI tool to control Unity Editor from the command line.
Unified successor to hera-agent + hera-agent-pro. All features ship free under Apache-2.0.
협업 개발 (Claude ↔ Codex)
hera-agent-unity는 Claude(Claude Code)와 Codex가 협업해서 개발하는 프로젝트다. 목표는 완벽한 hera를 함께 만드는 것 — 한 에이전트가 놓친 것을 다른 에이전트가 잡는다. (이건 hera를 사용하는 이야기가 아니라 hera라는 도구 자체를 개발하는 협업 원칙이다.)
- git 히스토리 = 공유 인계 채널. Codex는 대화 맥락이 아니라 git 커밋 히스토리로 프로젝트 상태를 파악한다. 그래서 작업은 명확한 conventional-commit 메시지로 커밋하고, 큰 기능은
docs/CODEX_HANDOFF_*.md같은 인계 문서를 남겨 다음 에이전트가 이어받게 한다. - 교차검증. 한쪽이 구현하면 다른 쪽이 리뷰·검증하는 것을 기본으로 한다. 자기 작업을 자기가 승인하지 않는다.
- 공통 규약 준수. 둘 다 이
CLAUDE.md(설계 의도·체크리스트·"이미 처리된 항목" 표)와AGENTS.md(hera 사용 규약)를 따르고, 🔒 잠긴 설계 결정을 존중한다. - 정확성 우선. 추측 대신 실측으로 근거를 만든다 —
hera-agent-unity로 라이브 Unity를 확인하고, 버전별 사실은 바이너리 리플렉션에서 뽑는다.
설계 의도
기존 Go CLI와 localhost HTTP Unity Connector가 실행 코어라는 결정은 유지한다 🔒. HttpServer, CommandRouter, ToolDiscovery, Heartbeat, Unity main-thread queue, 파일버스 복구 경로를 교체하거나 Unity Connector 안에 MCP를 직접 구현하지 않는다.
CLI + MCP adapter migration의 M0부터 M17까지 PASS다 🔒. 기존 바이너리 안의 선택적 Go stdio MCP adapter는 같은 실행 코어 앞에 있으며 Connector를 대체하거나, 도구 정의를 분기하거나, CLI 호환성을 제거하면 안 된다. CLI와 MCP는 하나의 정규화된 tool contract registry를 공유한다. M17에서 열네 개 section 28.3 시나리오와 복구 증거를 완결하고 독립 PASS B 승인을 받았다. Typed CLI와 기존 CLI가 production default이고 MCP는 experimental·default-off 상태를 유지한다. MCP를 명시적으로 켠 뒤의 exposure 기본값은 실측된 model-facing definition payload와 명시적 사용자 결정에 따라 Compact다. Profile과 Full은 큰 정적 schema payload를 감수하는 opt-in이다.
이유:
- 런타임 의존성 0개 — 사용자는 바이너리 하나 + UPM 패키지 하나만 설치
- Stateless — 모든 요청이 독립적이라 세션·재연결 로직 불필요
- 도메인 리로드를 파일시스템 버스(
~/.hera-agent-unity/instances/,status/)로 우회 - 어떤 셸·AI 에이전트·스크립트에서도 호출 가능 (MCP 클라이언트에 묶이지 않음)
MCP migration lock and ledger
- Authoritative implementation specification:
docs/CODEX_MCP_MIGRATION_IMPLEMENTATION.md - Milestone evidence and rollback ledger:
docs/MCP_MIGRATION_PROGRESS.md - 현재 상태: M0부터 M17까지 PASS다. CLI
v0.1.0+의mcp는HERA_MCP_ENABLED=1일 때만 stdio로 시작하며 Compact 3-tool default, opt-in Profile/Full-safe, 명시적 arbitrary-code permission이 필요한 Advanced, 승인/MRTR, operation ledger, Tasks와 blocking fallback, large-result resource를 지원한다. 오래된 Connector는 Compact-only로 보수적으로 저하되고 안전 feature 누락은 fail-closed다. 측정 이득 기준은 MCP 자체를 production default로 승격하지 않으므로 Typed CLI와 기존 CLI가 production default이고 MCP는 experimental·default-off다. - 보존 경계: 기존 Go CLI, localhost HTTP Connector, single-selected-target model, main-thread serialization, heartbeat discovery, package/test file bus, CLI/Connector 독립 버전은 계속 잠금 상태다.
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.
- 3d ago First seen · 542 lines · 11,102 tokens per session scan A 3c245def87ba
hera-agent-unity CLAUDE.md is an instructions file published in the GitHub repository NotNull92/hera-agent-unity (28 stars, last pushed 15d ago), licensed Apache-2.0. It adds 11,102 tokens to every session, about $0.0555 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 2026-08-30.
Other instructions, from other repositories
unity-cli-skill AGENTS.md
Instructions for niqibiao/unity-cli-skill, covering agents.md, project overview, invoking the cli, lifecycle and setup and command-first principle.
unity-cli-skill CLAUDE.md
Instructions for niqibiao/unity-cli-skill: Project instructions are maintained in and imported from @AGENTS.md.
open-agent-hub AGENTS.md
AGENTS.md instructions for guanyang/open-agent-hub, covering agents.md, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.
core AGENTS.md
Instructions for ReactUnity/core, covering release workflow, write changelog files, example, fix button hover state and package references.
gmgn-skills CLAUDE.md
Instructions for GMGNAI/gmgn-skills, covering claude.md, critical rule — read this first, project overview, available skills and quick decision guide.
core copilot-instructions.md
Instructions for asdecided/core: These decisions are already accepted. Do not re-open or contradict them; ask the AsDecided MCP tools (getartifact, searchartifacts) for the full text before proposing a change that touches one.