agentleFS
haxsdhaxsd/haxsd-byok/AGENTS.md

haxsd-byok

AGENTS.md33 starsChanged 33 days agoMITView on GitHub

本仓库对陌生人可见。内部资料——交接说明、内部方案与审查/审计记录、含机器路径或真实用量数据的笔记、 逆向工具——一律写进私有记录仓库 haxsd/haxsd-dev-records(本仓库对应分区 haxsd-byok/;历史分区 cursor-byok/ 对应的 GitHub 仓已删除), 公开仓库只放产品与贡献者需要的内容。 公开仓库不得出现这些路径:HANDOFF.md、BRANCH_ISOLATION.md、docs/plans/**、docs/superpowers/**、 docs/quality-review-*、docs/devin-router-vendor-evidence.md、support/cursor-capture/**、 support/cursor-protocol-extractor/**。CI 的 `reco

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,439 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository7 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

7 rules. About 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not preserve backward compatibility.

    line 3
  2. Distinctive:

    Keep components modular and concerns clearly separated.

    line 6
  3. Distinctive:

    Prefer established, well-maintained libraries when they reduce overall complexity or improve reliability.

    line 7
  4. Distinctive:

    use measured line counts for existing code.

    line 24
  5. Very specific:

    use clearly marked approximate targets such as ≈300 lines for proposed code.

    line 25
  6. Distinctive:

    Use the user's vocabulary consistently.

    line 29
  7. Distinctive:

    Prefer compact trees, flow diagrams, state matrices, and mappings when they communicate the structure more clearly than paragraphs.

    line 31

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

AGENTS.md

  • The directory structure is the architecture. Simple, clear directory and module naming >= module dependency relationships > concrete implementation details; communicate with the user using directory trees.
  • Do not preserve backward compatibility. Remove obsolete paths instead of adding compatibility layers, fallbacks, or migrations.
  • Choose the simplest implementation that fully meets the current requirements. Avoid speculative abstractions, configuration, and indirection.
  • Grow the system in layers. Start from the smallest version that works end to end, and add each new capability on top of a product that already works. Never trade a working product for unfinished complexity.
  • Keep components modular and concerns clearly separated.
  • Prefer established, well-maintained libraries when they reduce overall complexity or improve reliability. Do not reimplement common functionality without a clear reason.
  • Lean on the dependencies already in the project before writing your own implementation or adding packages. Do not assume a library lacks a capability without checking its documentation and types.
  • Make architectural decisions for the long term. Do not accept a stopgap that only works for now and is meant to be replaced later.

客户端行为变更前置

  • 修改 Cursor、Devin、Codex、Claude Desktop 的接管、协议、模型目录或客户端行为前,先读私有记录仓库 haxsd/haxsd-dev-records 中的 haxsd-byok/docs/reverse/README.md,再按 haxsd-byok/docs/product-change.md 定位对应版本的官方实现与本产品依赖。
  • 资料不可访问、版本不适用或证据不足时,先说明缺口并定向补证;不得用旧版偏移、字符串存在、第三方补丁或本产品测试结果替代官方实现证据。
  • 修复先复现并核实完整因果链,优化先确定基线和指标;再选择最小改动及回归范围,按用户可见结果验收。
  • 官方实现与版本证据回写私有逆向区;产品方案、代码映射、测试和交付状态写产品记录区。区分静态核对、自动化检查与真实客户端验证,不把未运行项记为通过。

Communication

  • Lead with the conclusion. Keep responses direct and omit filler, repeated context, generic explanations, and narration of obvious steps.
  • Match the level requested by the user: discuss architecture as architecture, behavior as behavior, and code details only when they materially support the answer.
  • For architecture, refactoring, and module-boundary discussions, communicate primarily with annotated directory trees and ASCII architecture/data-flow diagrams instead of long prose.
  • In directory trees, annotate every relevant directory and file with its single responsibility. When discussing code size, include line counts in the comments:
    • use measured line counts for existing code;
    • use clearly marked approximate targets such as ≈300 lines for proposed code;
    • state whether tests, generated code, and blank lines are excluded.
  • Show the complete main execution path, including inputs, ownership boundaries, runtime loops, persistence, outputs, and extension points. Make the direction of data and control flow explicit.
  • Clearly separate the current structure from the proposed structure. Explicitly list modules that move, merge, split, or are deleted.
  • Use the user's vocabulary consistently. Do not introduce new domain terms when existing plain-language terms are sufficient; when an implementation name must be mentioned, distinguish it from the architecture concept.
  • For stateful or concurrent behavior, show timing, ownership, state boundaries, and before/during/after behavior explicitly. Distinguish continuation within the same lifecycle from creation of a new lifecycle.
  • Prefer compact trees, flow diagrams, state matrices, and mappings when they communicate the structure more clearly than paragraphs.
  • Tie recommendations to the actual repository structure and code. Do not present a speculative target architecture as if it already exists.

仓库边界:内部资料不进公开仓库

本仓库对陌生人可见。内部资料——交接说明、内部方案与审查/审计记录、含机器路径或真实用量数据的笔记、 逆向工具——一律写进私有记录仓库 haxsd/haxsd-dev-records(本仓库对应分区 haxsd-byok/;历史分区 cursor-byok/ 对应的 GitHub 仓已删除), 公开仓库只放产品与贡献者需要的内容。

公开仓库不得出现这些路径:HANDOFF.md、BRANCH_ISOLATION.md、docs/plans/**、docs/superpowers/**、 docs/quality-review-*、docs/devin-router-vendor-evidence.md、support/cursor-capture/**、 support/cursor-protocol-extractor/**。CI 的 records-guard 任务会拦下它们。

提交信息里也别写内部细节(账号、机器路径、真实 token 数),这些写进记录仓库的提交信息。 收尾检查:在记录仓库里跑 sync-records.ps1(扫描五个公开仓库的当前树与历史,并检查开发区根目录只放两个产品仓库;发现漂移即报错)。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.