project-code-study 是一个面向真实代码项目的中文研究型学习 Skill。它把“读代码”组织成可复现、可追问、可校验的学习闭环：先从真实运行调用链建立路线，再一次只学习一个 RUN/NODE，通过主动回忆确认理解，最后把问答、进度、证据和学习文档持久化为可独立复习的材料。
长对话中的代码学习容易出现四类断裂：讲解脱离真实源码，模型把猜测说成事实；用户回答没有被逐项评价；QA/LOG/总结文件发生串写或互相矛盾；对话被压缩或中断后，学习状态丢失。这个 Skill 将这些问题拆成教学协议、持久化协议、证据核验和最终化门禁，并使用本地脚本进行严格检查。
它适合：
- 从真实运行调用链梳理深度学习、后端、工具链或其他复杂项目；
- 以“生活化例子 → 公式/规则 → 代码/语法 → 对应关系”的方式理解实现；
- 对输入输出、Shape、状态、配置、运行结果和论文对应关系进行证据化学习；
- 在长会话、上下文压缩或异常中断后继续学习，而不依赖模型记住全部历史。
它不是自动生成项目介绍的摘要器，也不是把源码逐行加注释的工具。完整代码讲解必须解释执行顺序、数据与状态变化、设计取舍、失败模式和证据边界；正式学习文档必须经过机器可验证的最终化流程。
flowchart TD
A[真实源码与运行证据] --> B[Preflight 建立 RUN/NODE 路线]
B --> C[讲解一个 NODE]
C --> D[主动回忆与复合问题拆分]
D --> E{回答是否完整且正确}
E -- 否 --> F[纠正 + retest-due]
F --> G{retest 通过}
G -- 否 --> F
G -- 是 --> H[记录事务：QA + 回读 + LOG + 对账]
E -- 是 --> H
H --> I{receipt 与 strict validation}
I -- 失败 --> J[unsaved-partial，阻断推进]
J --> H
I -- 通过 --> K[AWAITING_QUESTIONS_OR_CONTINUE]
K --> L[用户关闭问题并发出新的 continue]
L --> M[推进下一个 NODE]
M --> C
K --> N[生成 fresh readiness manifest]
N --> O{ready}
O -- 否 --> P[修复缺口]
P --> N
O -- 是 --> Q[schema 2.1 分层 Step 手册 + 真实冷启动]
Q --> R[统一 WAL + COMMITTED release receipt]
关键规则：
| 领域 | Skill 的强制约束 |
|---|---|
| 路线 | 基于真实运行调用链规划；单次只推进一个 RUN/NODE。 |
| 证据 | 声明按类型交给对应 verifier；源码、配置、运行、数学、论文、比较和学习者判断分别核验。 |
| 问答 | 用户问题进入回答和记录流程；复合问题拆成独立 Q-ID；错误或部分正确回答必须复测。 |
| 持久化 | Q/M/C/TX 由 allocator 分配；QA 写入、精确回读、LOG 更新、跨文件对账和 strict validation 组成事务。 |
| 状态 | 保存后停在 AWAITING_QUESTIONS_OR_CONTINUE；旧 continue、未保存问题和 retest-due 都不能推进主线。 |
| 失败处理 | 没有机器 receipt 不能声称 saved；部分失败返回 unsaved-partial 并 fail-closed。 |
| 最终化 | 正式文档必须是可检索、可跳读的逐 Step 手册；finalizer 只产生 release-pending，统一 release receipt 才能证明保存。 |
| 记忆 | 普通问题不入长期记忆；偏好、纠正、质量反馈和 Step 规则只生成 candidate，经批准和事务绑定后才能 saved。 |
v6.2 把使用方式收敛为一套固定模式。用户只有一个启动提示，之后直接自然
提问、回答回忆题、纠正或说“继续”；Skill 内部始终执行
定位 → 学习 → 检验 → 沉淀 → 等待。七个模式为 START、LEARN、
ASK、ASSESS、RECOVER、CLOSE 和 REPAIR，对应不同的短响应合同，
不再让启动、问答、恢复和修复机械套用同一套八段 NODE 模板。
一次消息可包含任意数量的独立问题。系统先生成带 source span/hash 的 input envelope，在回答前把所有问题按原顺序登记到 QA/LOG，并生成一个 intake receipt；随后每个原 Q-ID 用独立 TX 回答和校验。某一题失败只阻断当前题， 此前已回答题保持提交，后续题保持 pending。聊天默认每轮最多展示三题只是 阅读分页，不是队列上限。追问使用新 Q-ID 和 Parent Q；同一消息里的“继续” 在出现问题或纠正时立即过期。
| 产物/控制面 | 唯一职责 | 不能替代 |
|---|---|---|
PROJECT_STUDY_QA.md |
保存分类型、可独立阅读的完整问答；publication 模式执行 concept/code/shape/metric/review/correction 深度合同 | LOG 状态、memory、正式手册 |
PROJECT_STUDY_LOG.md |
当前 Step/RUN/NODE、主线锚点、retest、pending intents、证据与事务状态 | 完整教学答案 |
.project-study-memory/ |
只保存获批的 durable 偏好、纠正、项目规则和证据指针 | QA/LOG、聊天转录、源代码事实 |
PROJECT_STUDY_DOCUMENT.md |
每个完成 Step 的紧凑学习闭环、精选源码、检索索引、练习和答案 | 原聊天、完整源码副本或索引式摘要 |
release_transaction.py |
用一份 WAL/receipt 绑定四类产物、revision、readiness、validator、cold-start 和 exact response | 宿主未执行的 hook |
memory 状态固定为 candidate → approved → saved → stale，另有终态
rejected。自动候选触发条件是：明确长期教学偏好、用户纠正、输出/文档/
路线质量反馈、Step 完成后的 durable learning rule。拒绝后删除原内容，只留
M-ID、hash、状态和原因。压缩前 handoff 必须包含主线锚点、完成 NODE、开放
问题、pending intents、retest、最近 correction、证据、artifact hash 和唯一
下一行动；hash 不一致时进入 REPAIR_REQUIRED。
正式手册使用 schema 2.1，目标是“翻到一个 Step 就能复习”，不是“把聊天和源码
扩写成厚教材”。每个 Step 用 8 个固定槽位：30 秒定位、调用链与数据边界、
精选源码证据、核心机制、设计取舍与故障定位、项目例子与重要 QA、自测与参考
答案、证据边界与下一跳。它们继续覆盖问题、前置知识、RUN/NODE、I/O/Shape/
状态、公式、设计理由、错误表现和完成标准，但不再拆成 15 个等长小节。
每个条目声明 compact、standard 或 specialist 阅读层级，对应 450–1,200、
800–2,200、1,400–3,600 个非代码字符。源码摘录总预算分别为 24、60、120 行，
单段最多 45 行；源文件不少于 20 行时，同一 Step 最多引用 35%。80 字以上的
非代码段落不得跨 Step 原样重复。训练、Shape、指标或创新机制可在文档内唯一的
DEEP-DIVE-* 深讲一次，但每个 Step 仍须先给出本地核心答案。Step 4.x、6、10
强制使用 specialist。schema 2.0 可继续只读迁移审计；新的正式 publication
必须是 2.1。
顶部快速索引支持按 Step、关键词、源码/符号和 Q-ID 定位。重要 QA 的完整规范 答案只在最相关 Step 正文出现一次；顶层问题区只保留 Q-ID、Step、主题、一句话 结论和正文锚点，避免二次复制。冷启动也不再只测“是否能复述”，而是依次测试 能否定位、解释和完成应用题。
一次正常学习循环会产生两层输出：聊天中给出结论、关键原因、证据摘要、Q-ID、QA 位置和当前主线状态；QA 中保存可独立阅读的规范答案。学习产物包括：
| 产物 | 作用 |
|---|---|
PROJECT_STUDY_ROUTE.md |
RUN/NODE 路线、上下游、完成条件和证据边界。 |
PROJECT_STUDY_LOG.md |
已完成节点、问题状态、纠正、复测和事务回执。 |
PROJECT_STUDY_QA.md |
完整、可独立阅读的规范问答。 |
.project-study-memory/MEMORY.md |
可恢复的连续性记忆索引与当前恢复指针。 |
PROJECT_STUDY_DOCUMENT.md |
通过 readiness、紧凑手册 validator、真实检索/解释/应用冷启动和统一 release 后的正式学习手册。 |
推荐从以下入口理解实现：
- SKILL.md：运行时总协议和资源路由；
- user-prompts.md：用户意图路由，不把 QA、LOG、暂停和 readiness 门禁推回给用户；
- interaction-mode-protocol.md：七种模式、input envelope、任意问题队列和压缩恢复；
- prompt-workflow-patterns.md：将外部工作流抽象为通用提示词模式；
- continuity-memory-protocol.md：上下文压缩和异常恢复协议；
- teaching-output-contract.md：代码讲解和聊天/QA 双层输出契约；
- transaction-and-evidence-protocol.md：事务、receipt、校验和证据协议；
- scripts/：状态机、账本、最终化、记忆和声明门禁工具；
- tests/：旧测试与端到端回归测试。
将检出的整个目录复制或链接到宿主支持的 Skill 目录，并确保运行时可以执行 Python 3。安装位置取决于宿主的 Skill 加载规则，不应写死为某个特定项目或机器路径。
常见布局示例：
Claude Code 用户级：~/.claude/skills/project-code-study
Codex 用户级： ~/.codex/skills/project-code-study
项目级： <project>/<host-skill-directory>/project-code-study
请保留 skills/project-study-document 子目录。伴生的 project-study-document 负责将已验证的学习记录整理为 schema 2.1 分层 Step 手册；它不能绕过主 Skill 的事务和 readiness 门禁。
首次在某个项目中启用连续性记忆时，Skill 会先询问是否允许创建项目根目录下的 .project-study-memory/。只有在用户明确同意后才运行 sync_protocol_memory.py init ... --user-consent；拒绝或未回答都不会创建。
启动时只需一句自然语言；目标和基础可选：
我想开始学习这个项目。目标是 <读懂 / 复现 / 修改 / 研究扩展>，
我目前了解 <一句话基础>，希望重点学习 <可留空>。
正常继续时使用新的明确指令：
继续学习下一个 NODE。
之后可直接自然提问，问题数量不受协议限制；也可回答回忆题、纠正、暂停或发送新的“继续”。恢复与异常诊断见 user-prompts.md 的高级附录。用户不需要手工维护 Q-ID、QA、LOG、pending intents 或 receipt。
在提交或发布前运行：
python -m unittest discover -s tests -p "test_*.py" -v
python scripts/validate_learning_ledger.py <PROJECT_STUDY_LOG.md> --strict --publication --qa <PROJECT_STUDY_QA.md>
python scripts/validate_finalization_bundle.py --ledger <PROJECT_STUDY_LOG.md> --qa <PROJECT_STUDY_QA.md> --publication
python scripts/validate_protocol_memory.py <MEMORY_ROOT>
python scripts/cold_start_test.py --report <REPORT.json> --document <PROJECT_STUDY_DOCUMENT.md> --step <STEP> --handbook-schema 2.1
python skills/project-study-document/scripts/validate_study_document.py <PROJECT_STUDY_DOCUMENT.md> --ledger <LOG> --qa <QA> --repo-root <PROJECT_ROOT> --publication --cold-start-report <REPORT.json>
python scripts/release_transaction.py prepare --manifest <RELEASE_MANIFEST.json> --wal <RELEASE.wal.json> --response-file <RESPONSE.md>
python scripts/release_transaction.py commit --wal <RELEASE.wal.json> --receipt <RELEASE.receipt.json>
python scripts/response_claim_guard.py <RESPONSE.md> --receipt <RECEIPT.json>
git diff --check验证重点不是“validator 能报告错误”，而是验证错误后不存在成功旁路：正式目标文件不被创建或覆盖，没有 receipt 不能产生 saved，重复 ID、相邻 QA 污染、无效链接、占位路径、不完整 UNIT 和旧状态穿透都会被拒绝。
当前脚本能强制本地文件、状态、hash 和 receipt 约束，但不能凭 Skill 文本给宿主安装 pre-response 或真实 compact hook。真实宿主、不同模型、真实上下文压缩必须分别测试；未执行项写 not-run，静态 validator 通过不能替代宿主通过。
project-code-study/
├── SKILL.md
├── README.md
├── LICENSE
├── CHANGELOG.md
├── references/
├── scripts/
├── tests/
└── skills/
└── project-study-document/
- 只把项目源码、运行日志和用户明确纳入范围的材料作为输入；不要把密钥、令牌或隐私内容写入记忆、QA 或最终文档。
- 绝不把模型口头确认当作持久化成功；以 receipt、精确回读和 strict validation 为准。
- 发现证据不足时标记未知、请求补证或停止推进，不用猜测填空。
- 正式文档生成使用临时同目录文件、preflight、final validation 和原子替换；草稿只能标记为
incomplete-draft。 - 记录文件属于用户学习产物；Skill 的维护只修改协议、脚本、模板和测试，不覆盖已有学习记录。
本 Skill 借鉴了下列公开项目或文档中的通用思想，并进行了独立抽象和实现：
完整的仓库链接、核心实现方式、许可证、审计日活跃程度、适用性分类和明确 拒绝项见 GitHub 调研与致谢。 特别感谢 Engramory、Mem0、Letta、Zep Graphiti、LangGraph、OpenHands、 SWE-agent、Aider、AutoGen、CrewAI、learn-codebase、PocketFlow Tutorial Codebase Knowledge、RepoAgent、CodeTour、DeepWiki-Open、Diátaxis、 Material for MkDocs、mdBook、Rust by Example、Log4brains、MathTutorBench 和 EducationQ，以及 GitHub Copilot customization、Claude Code skills、 Anthropic skill-creator、Superpowers 和 awesome-copilot 的维护者与研究者 公开相关思想。
| 项目/文档 | 借鉴的通用思想 | 说明 |
|---|---|---|
| Engramory | 文件化连续性记忆、受控索引、单事实记录、去重/更新/归档 | 启发了本 Skill 的 memory 协议；未复制其源代码。该仓库 README 标注为 MIT。 |
| CodeTour | 有序代码导览、文件/行选择、primary tour 与下一跳 | 启发了 Step 顺序、精选源码锚点和前后导航；未复制其扩展代码。 |
| Diátaxis | tutorial/how-to/reference/explanation 分工和按需深入 | 启发了“Step 核心闭环 + 检索索引 + 共享深讲”的分层结构；其文档为 CC-BY-SA 4.0，本仓库只借鉴思想。 |
| Material for MkDocs 与 mdBook | 搜索、目录、锚点、前后导航和源码归属 | 启发了单 Markdown 内的快速索引；未引入站点运行时。 |
| Rust by Example | 小而完整的例子和可应用练习 | 启发了每 Step 一个项目最小例子；未复制示例内容。 |
| Log4brains | 轻量 Markdown、可搜索元数据和渐进披露 | 启发了内容预算、去重和可选深读；未复制其实现。 |
| learn-codebase | 苏格拉底式提问、先预测后揭示、主动回忆、渐进式支架和学习日志 | 启发了主动回忆、复测和教学输出契约；许可证以其仓库为准。 |
| VS Code Agents documentation | Understand → Act → Validate 循环、计划阶段、动作结果反馈 | 启发了本 Skill 的执行—验证闭环；许可证以其仓库为准。 |
| GitHub Awesome Copilot | preflight、条件步骤、失败即停、结构化包装提示和输出契约 | 启发了提示词路由与 fail-closed 规则；许可证以其仓库为准。 |
| Superpowers | 计划检查点、分步执行、验证门和阻塞时停止 | 启发了节点推进和最终化前检查；该仓库标注为 MIT。 |
| AGENT.md specification | 分层指令、作用域继承和可预测的项目上下文 | 启发了资源分层与恢复入口；许可证以其仓库为准。 |
| GitHub Copilot onboarding plan | Foundation → Exploration → Integration 的分阶段学习路径 | 启发了启动、探索、整合的提示词组织方式。 |
| GitHub Copilot customization、Claude Code skills 与 Anthropic skill-creator | 用户输入、Skill、按需资源和宿主 hook 的职责分层；渐进披露 | 启发了 v6.2 单启动入口、短合同和协议资源分层；未复制提示词或协议。 |
本项目与上述项目没有隶属、赞助或背书关系。第三方项目的版权和许可证均归其各自权利人所有；使用或再分发第三方代码、文本或资产时，应直接遵守对应仓库的 LICENSE、版权声明和贡献者要求。本仓库当前仅借鉴公开的工作流思想，未将上述项目的源代码、提示词原文或资产作为本 Skill 的组成部分。
本项目采用 MIT License。
project-code-study is a Chinese-first, evidence-bound Agent Skill for studying real software projects. It turns a verified runtime call chain into a route of RUN/NODE units, teaches one node at a time, evaluates active recall, and persists questions, progress, evidence, and study documents for independent review.
Version 6.2 adds one natural start prompt, the standard
locate → learn → assess → persist → wait loop, seven response modes, a
source-bound intent envelope, and arbitrary-size question batches that register
every Q before answering and commit each answer independently. It retains the
compact schema 2.1 Step-manual entries, quick lookup indexes,
reading and source-excerpt budgets, cross-Step duplication checks, document-local
deep dives, and retrieval/explanation/application cold-start reports. It keeps
the v6 typed durable-memory candidates, hash-bound compaction handoffs,
type-specific QA depth contracts, exact source evidence, and one WAL-backed
release receipt binding QA, LOG, memory, document, validators, source revision,
not-run boundaries, and the exact response. See the
research and acknowledgements table
for upstream ideas, licenses, activity, adoption decisions, and explicit
non-adoptions.
Long code-learning conversations can drift away from the real source, turn guesses into facts, lose learner answers, corrupt adjacent QA entries, or confuse a conversational acknowledgement with a successful write. This Skill addresses those failure modes with an executable teaching protocol, persistence transactions, claim verification, continuity memory, and finalization gates.
It is suitable for learning deep-learning, backend, tooling, and other complex repositories from their real execution paths. It is not a generic project-summary generator: copying source code and adding short comments does not qualify as a complete explanation.
Verified source/runtime evidence
-> preflight and RUN/NODE route
-> teach exactly one NODE
-> split compound questions and run active recall
-> evaluate, correct, and retest when needed
-> commit QA + readback + LOG + reconciliation as one transaction
-> require receipt and strict validation
-> wait in AWAITING_QUESTIONS_OR_CONTINUE
-> advance only after a new continue
-> build a fresh readiness manifest
-> build schema 2.1 layered Step-manual entries
-> run a real fresh-model/document-only lookup, explanation, and application cold-start
-> stage through the finalizer
-> commit one hash-bound release receipt
| Area | Enforced behavior |
|---|---|
| Route | Build the route from a real runtime call chain and advance one RUN/NODE at a time. |
| Evidence | Route source, configuration, runtime, mathematical, paper, comparison, and learner-verdict claims to the appropriate verifier. |
| Questions | Enter the answer-and-record flow after a learner question; split compound intents into independent Q-IDs; retest incorrect or partial answers. |
| Question batches | Register every source-bound Q before the first answer; update one existing Q per TX; preserve earlier commits and later pending Qs after a failure. |
| Persistence | Allocate Q/M/C/TX IDs uniquely; combine QA write, exact readback, LOG update, reconciliation, and strict validation into a transaction. |
| State | Stop at AWAITING_QUESTIONS_OR_CONTINUE; unsaved questions, stale continue tokens, and retest-due states cannot advance the main route. |
| Failure | Do not claim saved without a COMMITTED release receipt bound to the exact response; return unsaved-partial or release-pending. |
| Finalization | Require schema 2.1 compact entries, selected exact source excerpts, real retrieval/explanation/application cold-start, and one release receipt; preserve the target when ready=false. |
| Memory | Create candidates only for durable preferences, corrections, quality feedback, and Step rules; approve/reject explicitly and restore from hash-bound handoffs. |
The final document is a manual to consult, not a textbook dump. Each Step uses
eight slots: quick orientation; call/data boundary; selected source evidence;
core mechanism; trade-offs and failure diagnosis; project example and selected
QA; self-test and answer; evidence boundary and next hop. compact, standard,
and specialist profiles cap non-code prose at 1,200, 2,200, and 3,600
characters and total excerpts at 24, 60, and 120 lines. One excerpt is at most
45 lines, a long source file may be quoted at most 35%, and long prose
paragraphs cannot be copied across Steps. Shared mechanisms live once under a
document-local DEEP-DIVE-*; each Step still contains its local answer.
The normal loop produces two layers of output. Chat provides the conclusion, key reasons, evidence summary, Q-ID, QA location, and current route state. QA stores the complete, independently readable canonical answer.
| Artifact | Purpose |
|---|---|
PROJECT_STUDY_ROUTE.md |
Route, upstream/downstream nodes, completion conditions, and evidence boundaries. |
PROJECT_STUDY_LOG.md |
Completed nodes, question states, corrections, retests, and transaction receipts. |
PROJECT_STUDY_QA.md |
Complete, independently readable canonical answers. |
.project-study-memory/MEMORY.md |
Recoverable continuity-memory index and resume pointers. |
PROJECT_STUDY_DOCUMENT.md |
Formal per-Step manual released after readiness, schema 2.1 compactness/source/navigation validation, real cold-start, and one committed receipt. |
Start with SKILL.md. The learner uses the single start prompt and natural follow-ups in references/user-prompts.md; interaction-mode-protocol.md defines the seven modes and arbitrary question queue; consult the remaining protocol references and scripts/ for the executable control plane.
Copy or link the complete checked-out directory into a Skill directory supported by the host. The destination is host-specific; do not hard-code a path from one project or machine.
Typical layouts are:
Claude Code user scope: ~/.claude/skills/project-code-study
Codex user scope: ~/.codex/skills/project-code-study
Project scope: <project>/<host-skill-directory>/project-code-study
Keep the skills/project-study-document companion directory. It turns validated study records into searchable schema 2.1 Step-manual entries and cannot bypass the main Skill's transaction, cold-start, or release gates.
When continuity memory is first enabled for a project, the Skill asks for explicit consent before creating <PROJECT_ROOT>/.project-study-memory/. A decline or missing answer creates no memory directory. The initialization command requires --user-consent.
I want to start learning this project. My goal is <understand / reproduce /
modify / research-extend>; my current background is <optional>.
To continue normally:
Continue with the next NODE.
After that, ask any number of questions naturally, answer recall, correct a claim, pause, or send a fresh continue. The Skill—not the learner—maintains Q IDs, QA/LOG, queues, receipts, recovery, and readiness gates.
Run the following before publishing:
python -m unittest discover -s tests -p "test_*.py" -v
python scripts/validate_skill_structure.py
python scripts/validate_learning_ledger.py <PROJECT_STUDY_LOG.md> --strict --publication --qa <PROJECT_STUDY_QA.md>
python scripts/validate_finalization_bundle.py --ledger <PROJECT_STUDY_LOG.md> --qa <PROJECT_STUDY_QA.md> --publication
python scripts/validate_protocol_memory.py <MEMORY_ROOT>
python scripts/cold_start_test.py --report <REPORT.json> --document <PROJECT_STUDY_DOCUMENT.md> --step <STEP> --handbook-schema 2.1
python skills/project-study-document/scripts/validate_study_document.py <DOCUMENT> --ledger <LOG> --qa <QA> --repo-root <PROJECT_ROOT> --publication --cold-start-report <REPORT.json>
python scripts/release_transaction.py prepare --manifest <MANIFEST.json> --wal <WAL.json> --response-file <RESPONSE.md>
python scripts/release_transaction.py commit --wal <WAL.json> --receipt <RECEIPT.json>
python scripts/response_claim_guard.py <RESPONSE.md> --receipt <RECEIPT.json>
git diff --checkValidation must also prove that no success bypass exists after a failure: the formal target remains unchanged, saved requires one committed release receipt, duplicate IDs and boundary pollution are rejected, invalid or over-budget source excerpts fail, shallow or bloated entries cannot validate, repeated long paragraphs fail, and stale state cannot advance the route.
The repository contains local/static regression tests. Real host behavior, cross-model behavior, and cross-session persistence must be tested in the target host separately and must not be reported as passing based only on a static proxy.
project-code-study/
├── SKILL.md
├── README.md
├── LICENSE
├── CHANGELOG.md
├── references/
├── scripts/
├── tests/
└── skills/
└── project-study-document/
- Treat only repository source, runtime logs, and explicitly scoped materials as inputs; never write secrets or tokens into memory, QA, or final documents.
- Treat receipts, exact readback, and strict validation—not model acknowledgement—as proof of persistence.
- Mark unknowns and request evidence when claims cannot be verified; do not fill gaps with guesses.
- Use temporary same-directory files, preflight, final validation, and atomic replacement for formal output. Drafts must be marked
incomplete-draft. - Treat learner records as user-owned artifacts. Maintain the Skill by changing protocols, scripts, templates, and tests rather than overwriting study records.
This Skill independently abstracts workflow ideas from the following public projects and documents:
| Project/document | General idea referenced | Note |
|---|---|---|
| Engramory | File-based continuity memory, bounded indexes, one-fact records, deduplication, update, and archive discipline. | Inspired the memory protocol; no source code was copied. Its README identifies the project as MIT-licensed. |
| CodeTour | Ordered code tours, file/line selections, primary tours, and next-step navigation. | Inspired Step ordering, selected source anchors, and continuation links; no extension code was copied. |
| Diátaxis | Separation of tutorial, how-to, reference, and explanation, with deeper material available on demand. | Inspired the local Step closure, lookup index, and shared deep-dive layers; ideas only. |
| Material for MkDocs and mdBook | Search, tables of contents, anchors, breadcrumbs, and previous/next navigation. | Inspired navigation inside one Markdown artifact; no site runtime was added. |
| Rust by Example | Small, complete examples and application-oriented learning. | Inspired one minimal project example per Step; no examples were copied. |
| Log4brains | Lightweight searchable Markdown knowledge and progressive disclosure. | Inspired content budgets, deduplication, and optional deeper reading; no implementation was copied. |
| learn-codebase | Socratic questioning, prediction before reveal, active recall, graduated scaffolding, and learning journals. | Inspired the recall, retest, and teaching contracts; see the upstream repository for its license. |
| VS Code Agents documentation | Understand → Act → Validate loops, plan phases, and action-result feedback. | Inspired the execution/validation loop; see the upstream repository for its license. |
| GitHub Awesome Copilot | Preflight, conditional steps, fail-stop behavior, structured wrappers, and output contracts. | Inspired prompt routing and fail-closed rules; see the upstream repository for its license. |
| Superpowers | Plan checkpoints, stepwise execution, verification gates, and stopping on blockers. | Inspired node advancement and finalization checks; its repository identifies the project as MIT-licensed. |
| AGENT.md specification | Layered instructions, scope inheritance, and predictable project context. | Inspired resource layering and recovery entry points; see the upstream repository for its license. |
| GitHub Copilot onboarding plan | Foundation → Exploration → Integration learning phases. | Inspired the organization of startup, exploration, and integration prompts. |
| GitHub Copilot customization, Claude Code skills, and Anthropic skill-creator | Separation of user input, reusable Skill workflows, scoped resources, and host hooks; progressive disclosure. | Inspired v6.2's single entry point, short mode contracts, and layered protocols; no prompt or protocol text was copied. |
This project is not affiliated with, sponsored by, or endorsed by the projects above. Copyright and licensing remain with each respective rightsholder. If third-party code, text, or assets are later copied or redistributed, follow the corresponding upstream LICENSE, copyright notices, and contributor requirements. This repository currently claims only independent reuse of public workflow ideas, not upstream source code, prompt text, or assets.
This Skill is released under the MIT License.
See CHANGELOG.md for releases.