TheColliny GitHub avatar

FableClaudeMDForOpus

TheColliny

Guardrails Kit v1.0 是一套可移植的 CLAUDE.md 和文档集,通过显式、可检查的流程让 Claude Opus/Sonnet 在 Claude Code 中减少逻辑错误、引入的 bug 和令牌浪费。

Stars

243

7 天增长

暂无数据

Fork 数

44

开放 Issue

0

开源协议

暂无数据

最近更新

2026-07-05

AI 仓库情报摘要
FR-AI / ANALYSIS

为什么值得关注

它将模型隐式推理转化为显式的、事件触发的步骤,通过精益的常驻规则和丰富的按需手册,使较弱模型能接近前沿水平,并通过可 grep 的标记验证合规性。

适合谁使用

  • 使用 Claude Code 进行编码的开发者
  • 将 AI 助手集成到软件开发流程中的团队
  • 希望最大限度减少 AI 生成错误和令牌浪费的高级用户

典型使用场景

  • 提高 AI 生成代码的质量并减少 bug
  • 跨会话保持一致的规划和验证流程
  • 通过结构化日志和检查点调试和跟踪 AI 推理过程

项目优势

  • 核心仅约45行,配合按需文档,令牌开销低但覆盖全面
  • 事件短语路由确保相关文档在需要时被读取,而非按主题标签
  • 基于数字的阈值(如“10条消息”、“>50次命中”)实现机械性合规而非判断
  • 每个禁止项都附带替代方案,防止规避行为

使用前须知

  • 需要手动安装和针对每个项目配置
  • 效果取决于模型遵循显式规则和标记的能力
  • 对于简单任务可能引入不必要的开销

README 快速开始

Guardrails Kit v1.0

A portable CLAUDE.md + documentation set that makes Claude Opus / Sonnet operate as close to frontier (Fable) level as possible inside Claude Code: fewer logic errors, fewer introduced bugs, fewer wasted tokens. It works by converting the implicit judgment a stronger model applies automatically into explicit, checkable, event-triggered procedures a weaker model can execute mechanically.

What's in the kit

FileRole
CLAUDE.mdAlways-loaded core: 12 iron rules, an event-phrased routing table, 4 hard stops. Deliberately small (~45 kit lines, hard cap 60) — always-on compliance is roughly constant-sum, so every extra line taxes obedience to all the others.
docs/guardrails/PLAN.mdBefore starting non-trivial work: TASK block, premise check, prior-art search, baseline, decomposition, ask-vs-decide.
docs/guardrails/CODE.mdWhile editing: read-before-edit gates, twin/generated-file checks, and the REFERENCE SWEEP procedure (RS1-RS5).
docs/guardrails/TRAPS.mdLookup tables for the classic reasoning traps: dates, epochs, mutation-vs-copy, async, floats/money, sort, division/modulo, regex, familiar-API lookalikes, closures, boolean logic. Read on demand via CODE.md C7.
docs/guardrails/DEBUG.mdWhen anything fails: reproduce-first loop, CAUSE line, failed-attempts ledger with the ESCALATION LADDER, red-flag table keyed on the model's own rationalization phrases.
docs/guardrails/VERIFY.mdBefore claiming done/committing: 12-item echo protocol — every claim needs output quoted from a real tool result in the same turn.
docs/guardrails/EFFICIENCY.mdToken/context discipline as paired rules: every "read less" rule has a "read enough" floor.
docs/guardrails/SESSION.mdLong-session survival: docs/STATE.md template (S2), same-turn update triggers (S3), post-compaction recovery (S1), ANCHOR/DETOUR/DECISION ledger keywords.
docs/guardrails/_FORMAT.mdAuthoring contracts for editing the kit itself (budgets, trigger phrasing, single-sourcing, sanctioned iron-rule pairs).
MIGRATE.mdThe transport procedure for retrofitting a project that already has a CLAUDE.md — line-accounted, backup-first, verbatim-carry, user-checkpointed, idempotent, with an UPGRADE mode.

Not part of the installable kit (kit-source materials only): docs/research-digest.md (the 155-findin

项目描述

A Fable written Claude MD for Opus/Sonnet to give it some of the quality methodology that Fable uses

相关仓库与替代方案

根据分类、Topic 和编程语言匹配的相似项目。

MoonshotAI
精选
MoonshotAI GitHub avatar

Kimi-K3

Kimi K3 is an open-weight, 2.8T-parameter native multimodal agentic model with a 1M-token context window, designed for frontier coding, knowledge work, and reasoning tasks.

AI 与机器学习AI 智能体
3,348
xuchonglang
精选
xuchonglang GitHub avatar

investing-for-beginners

A structured investing guide for Chinese beginners covering US stocks, options, and cryptocurrency, with focus on foundational concepts and risk awareness.

区块链与 Web3
2,739
Krishnagangwal
精选
Krishnagangwal GitHub avatar

CS-Fundamentals

A curated collection of Computer Science fundamentals (PDFs, notes, cheatsheets, interview question banks) for placement preparation, covering seven core subjects plus general resources.

数据与数据库数据库与存储
2,326