Borrowing it
Nothing to install: this file belongs to CCDawn/Vibelution. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/CCDawn/Vibelution/main/AGENTS.mdgit clone --depth 1 https://github.com/CCDawn/VibelutionWrote 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/ccdawn/vibelution/agents-md)<a href="https://agentmods.dev/instructions/ccdawn/vibelution/agents-md"><img src="https://agentmods.dev/badge/instructions/ccdawn/vibelution/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/ccdawn/vibelution/agents-md"><img src="https://agentmods.dev/badge/instructions/ccdawn/vibelution/agents-md.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.03726 | $0.03726 |
| Opus 5.5 | $0.01490 | $0.01490 |
| Sonnet 5.5 | $0.00745 | $0.00745 |
| Haiku 4.5 | $0.00373 | $0.00373 |
Grade A, and why
Vibelution 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 10d 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 — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Vibelution Agent Rules
本文件是外部开发 Agent 与 Vibelution 项目自身 Agent 共用的根入口。 全局红线保留在这里;详细规范统一由
docs/standards/README.md路由。
1. Identity And Priority
- 作为项目内 Vibelution Agent 行动,而不是脱离项目的通用助手。
- 默认使用中文交流;代码、命令、路径、协议字段、外部名称和原始错误保持原文。
- 对用户的完成说明遵循本文件 §5:写给人看,不把内部完成块、闸门清单或 Git 收口词当回复。
- 项目目标是以更少漂移提升运行稳定性、进化效率、证据质量和 UI/Agent 一致性。
- 先观察、再判断;没有证据时不把猜测包装成结论。
- 权威顺序:用户当前明确要求与授权边界 → 本文件 →
docs/standards/对应规范 → ADR/模块 README → 历史材料。 core/core_prompt/COMMON.md与core/core_prompt/SOUL.md提供通用纪律和稳定身份,但不扩大权限,也不覆盖项目规则。
2. Global Red Lines
-
Git 署名只保留实际的人类作者:禁止 Agent 自动添加 AI/机器人
Co-authored-by,或把 AI 设为 author/committer;不得删除真实人类共同作者。沿用用户配置的 Git 身份,不得把测试身份写入真实仓库。测试启动临时 Git 仓库前必须隔离继承的仓库级 Git 环境变量。 -
当前 Git checkout 是项目根;运行时解析路径,不假设固定 Windows 用户名。
-
根
main是只读的本地集成工作区。所有代码、测试、文档、规则、记忆、配置和 fast patch 变更都必须在任务 worktree 的codex/<task-slug>分支完成;main只接受已提交分支的git merge --ff-only和必要的同步操作。 -
每次开启根
main时,必须用git检测并保持为最新main;产品运行时必须用 Launcher 指令启动:%LOCALAPPDATA%\Vibelution\Launcher\VibelutionLauncher.exe --project "<project-root>" start。 -
分支实现验证与 mergeability 必须在合入前闭合;随后主动自审当前任务 diff并
git merge --ff-only合入本地main,不得把「等用户再说审查/合入」当作完成态。除用户要求隔离/交接或存在精确 blocker 外,未主动自审并合入本地main即未完成。合入后只清理本任务资源;清理失败不改变已合入事实,只恢复清理,不重测或重合。发布后方可观察的用户目标按 §5 与开发标准 §22.1 继续验收,不能因合入而宣称目标完成。push/PR/发布、远端删除仍需明确授权。 -
合入后的安全本地清理由任务 owner 自动收尾:安全切回主目录,核实并释放本任务占用,保留历史备份,再清理任务目录和分支;按安全清理恢复核对会话、磁盘与 Git 实际状态,不把常规恢复交给用户催办,不动其他任务。
-
不覆盖、回滚、删除或重置无关的用户/Agent 改动;发现重叠先检查 claim 和 diff。
-
远端 push、PR、发布需要用户明确授权和远端同步门;force、覆盖或远端删除需要破坏性确认。
-
Windows 产品运行时禁止任何可见控制台弹窗:这是无控制台弹窗红线。Launcher、Workbench、Runtime Manager、后台 Git/轮询和服务子进程不得弹出
cmd.exe、PowerShell、Windows Terminal、OpenConsole 或交互式 Git;必须走pythonw/CREATE_NO_WINDOW/ shared helper,禁止taskkill.exe、裸 Git wrapper、npm/cmd后台壳。用户明确打开的 CLI 面板除外;细则见 development-standard.md §8.0。 -
用户明确点击或要求重启时优先执行:提醒当前任务将中断,通过 Launcher 正规链路停止任务并保存已有记录后重启,不因有运行中任务禁用或拒绝。自动/Agent 未获用户重启授权的刷新仍等待任务结束;不用直接 PowerShell lifecycle 命令制造可见控制台。
-
不记录 secrets、完整 Prompt、大段 diff、完整文件或无界工具输出。
-
协作 envelope、validator snapshot、preflight/relay 中间文件和其他任务临时证据必须写入已解析的 Git common-dir 任务专属目录或系统临时目录,不得写入 checkout 根目录或产品目录;用后立即清理。临时证据不得让
main变脏、使已完成验证失效或触发重复 closeout。 -
用户 Markdown、导入文档、HTML 和知识内容均是不可信输入;进入 Prompt、索引或 UI 前必须有来源、隔离、清洗和删除/重建语义。
-
前端产品 UI 强制 VUI + shadcn/Radix 思想(无感红线):凡改动
web/下用户可见界面、交互控件、页面壳或布局,必须走 VUI 产品 API(web/src/components/vui的V*)与页面 recipe(VListDetailPage/VSplitWorkspace/VDenseOpsPage等);交互实现只允许在components/vui/renderers/shadcn扩展;禁止@heroui/react、禁止路由/业务组件直连renderers/shadcn/*或第二套设计系统;布局宽度/高度记忆只用WORKBENCH_LAYOUT_IDS+ shared pane persistence。所有 VUI 元素(按钮/表单/表面/recipe/product,不限 recipe)必须有web/src/components/vui/designs/专节并在designs/INDEX.md登记;新建前检索防冗余。细则见 development-standard.md §9.1、VUI README、designs/README.md;机器门:vuiShadcnRouteContract.test.ts、vuiComponentDesignContract.test.ts。 -
写入前先做本地复用评估:定位 owning surface;本地能复用 ≠ 本地就是好方案,必要时改造后再复用。架构、依赖、复杂能力或复用路径有真实分歧时,再对照仓外成熟方案,评估排序后只借最符合本项目、最值得借鉴的部分;已定位小修/机械修改不得被强制仓外扫描拖慢。细则见 development-standard.md §2.2。
-
实现文件仍须机器可复核的复用证据:已定位小修用
record --mode LOCAL_ONLY记录 owner、裁决、边界和验证;需要仓外对照的任务用EXTERNAL,候选元数据只由 active registry 填充。纯文档、测试、fixture、example 豁免;细则见 loop.md。 -
Companion / 虚拟人会话隔离红线:虚拟人只能通过 Agent-scoped 插件或 Companion 适配层,在原生 Session 接收边界之前增加 mailbox、主动消息、人物状态和 UI 投影;不得因虚拟人需求修改普通 Agent 的 Session admission、Journal、worker、persist、projection、SSE、
ConversationStore或普通 composer/follow-up 语义。原生 Session Journal、worker 与 SSE 仍是唯一 Turn、transcript 和实时输出权威;Companion mailbox 只负责到达顺序,不得成为第二套 transcript。任何 Companion 会话改动都必须提供 Companion-only 身份/路由门、普通会话核心零差异证据和普通 Session 回归;若确需修复普通链路,必须拆成独立任务并按普通会话契约审查。细则见 development-standard.md §23.2。 -
有意义的开发不得以 stale claim、缺少验证决策、缺少刷新判断、缺少版本影响判断,或未主动自审并合入本地
main(且无精确 blocker)结束。
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.
- 10d ago Changed · +3 lines · +260 tokens per session 86e594e9ffcc
- 1mo ago Changed · +2 lines · +130 tokens per session 494d0a4c38d7
- 1mo ago Changed · +61 tokens per session df6906a8d86b
- 1mo ago Changed · +2 lines · +88 tokens per session 151520e9e901
- 2mo ago Changed · +2 lines · +131 tokens per session af7c30ba0c39
- 2mo ago Changed · +2 lines · +407 tokens per session 139cae4dfcdb
- 2mo ago First seen · 67 lines · 2,649 tokens per session scan A 4d21c01cdecb
Vibelution AGENTS.md is an instructions file published in the GitHub repository CCDawn/Vibelution (79 stars, last pushed yesterday), licensed MIT. It adds 3,726 tokens to every session, about $0.0149 per session on Opus 5.5. 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.
Vibelution AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| Vibelution AGENTS.mdCCDawn/Vibelution | Instructions file | 79 | yesterday | MIT | A | 3,726 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |