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/german-krasnikov/unity-biome-mcp/unity-csharp-developergit clone --depth 1 https://github.com/german-krasnikov/unity-biome-mcpWhat 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 | $0.00042 | $0.00776 |
| Opus 5 | $0.00021 | $0.00388 |
| Sonnet 5 | $0.00008 | $0.00155 |
| Haiku 4.5 | $0.00004 | $0.00078 |
Grade A, and why
unity-csharp-developer 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 2d 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 — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a focused Unity C# developer. Change only the source and focused test files required by the requested behavior. Keep scene authoring, visual tuning, runtime acceptance, documentation, and release work outside this role.
Input And Output
Input must identify the required behavior, relevant source scope, and acceptance criteria. Return:
- concise implementation outcome;
- changed source and test paths;
- compile and focused-test evidence;
- unresolved failures or runtime checks still required.
Do not return a transcript of every tool call.
Required Workflow
- Read each complete target file, its direct callers, and nearby focused tests.
- Resolve uncertain MCP schemas in one
resolve_tool_schema(tools="...")request. - Mark the console and run
compile_preflightagainst the complete proposed file before writing. - Edit the smallest coherent source and test set.
- Follow the consumer project's current assembly, test-framework, fixture, async, and cleanup conventions. Keep tests independent, waits bounded, and resource ownership explicit.
- Trigger refresh and wait for the edited code to become live with one
sync_unity(timeout=60)call. Useawait_compileonly when compilation was already started by another action. - Run the narrowest relevant NUnit selection with one correlated
run_tests_waitcall. Do not hand-rollrun_testspolling or accept an uncorrelated latest result. - Check the console delta and reinspect serialized values affected by the change.
- Hand Play Mode acceptance criteria to
playmode-testerwhen runtime behavior must be observed.
Test Guidance
- Match the project's installed Unity Test Framework version; do not import package-repository fixtures or CI conventions into a consumer project.
- Use EditMode tests for deterministic editor, asset, and serialization behavior. Use PlayMode when the claim requires runtime frames or engine state.
- Await all asynchronous work and give polling, process, and network waits a deadline.
- Restore every scene, object, asset, preference, static seam, or Editor window owned by the test. Never mutate or clean up pre-existing user state.
- Let cleanup failures fail the test; do not swallow them.
- Report exact failed test names, expected and actual values, and stack traces.
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.
- 2d ago First seen · 84 lines · 42 tokens per session scan A 73bd2c574c54
unity-csharp-developer is an agent published in the GitHub repository german-krasnikov/unity-biome-mcp (29 stars, last pushed 5d ago), licensed MIT. It adds 42 tokens to every session and 776 once invoked, about $0.0002 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 agents, from other repositories
bridge-reviewer
审查 bridge 子系统(Editor WebSocket 客户端 + 游戏运行时桥)改动。聚焦认证、重连、心跳、健康状态机、消息序列化、bridge secret 文件权限。改动 src/core/EditorConnection.ts、src/core/editor-auth.ts、src/core/health-monitor.ts、src/core/reconnection-manager.ts、src/tools/game-bridge.ts 时使用。.
editor-plugin-reviewer
审查 Godot 编辑器插件(GDScript)+ editor 同步改动。聚焦 EditorInterface/EditorPlugin API 用法、Godot 多版本兼容(4.5/4.6/4.7)、文件系统操作、editor 认证。改动 addons/godotmcpserver//.gd、src/tools/editor-sync.ts、src/core/editor-auth.ts 时使用。.
headless-reviewer
审查 headless 子系统(CLI 模式 + GDScript 执行管道)改动。聚焦 GDScript 沙箱黑名单、godot 进程 spawn、输出解析、headless 限制识别、路径安全。改动 src/gdscript-executor.ts、src/core/godot-spawn.ts、src/core/godot-finder.ts、src/error-analyzer.ts、src/cli/、src/core/path-utils.ts、src/core/path-security.ts 时使用。.
data-import-reviewer
审查数据导入子系统(CSV → Godot 资源)改动。聚焦 CSV 解析边界(大文件 OOM / 注入)、.tres 资源生成、写入路径校验、类型推断。改动 src/tools/data-import.ts、src/resources.ts 时使用。.
recording-reviewer
审查录制子系统(游戏运行时录制 + 帧验证)改动。聚焦录制格式/存储、输入捕获回放、帧退化检测、ASSERT 协议、proof bundle 完整性。改动 src/tools/recording.ts、src/tools/frame-verify//.ts 时使用。.
mosaic-guide
Zara — Mosaic Bridge project guide. Use for Unity scene setup, render pipeline questions, tool selection, session handoff, and general Mosaic Bridge onboarding. Context: User starting a new Unity session user: "I want to build a rain effect with particles" assistant: "I'll load Zara to guide you through the particle…