agentsclimarketplace

Towow dev handoff

Skill floccose-burner9185/wow-harness/reference/towow-snapshot/.claude/skills/towow-dev-handoff

Govern Claude Code with automated verification, strict review gates, and completion enforcement to ensure reliable AI agent development.

Install
npx -y skills add floccose-burner9185/wow-harness --skill towow-dev-handoff

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

通爻开发接手入口。用于新 AI 接手 Towow 开发、梳理最新上下文、开始生产排障、继续上一轮工作或确认当前主线时使用。

SKILL.md

5.1 KB, ~1.6k tokens by cl100k_base, as published. Nobody here has run it

Towow 开发接手入口

目标

这个 skill 不负责复述 Towow 的全部历史，也不负责重新总结理念。

它只负责一件事：让新接手 Towow 开发的 AI，在一次入场中找到最新、可执行、可验证的上下文入口。

默认场景：

  • 接手 Towow 生产开发
  • 梳理当前主线
  • 开始生产排障
  • 继续上一轮工作
  • 进入 MCP / protocol / matching / run 链路的真实代码

启动协议

每次新 AI 接手 Towow 时，按这个顺序执行：

  1. 先运行： 
    python3 .claude/skills/towow-dev-handoff/scripts/collect_handoff_context.py
  2. 先读 repo 真相源，再读派生资料。
    • 如果任务涉及公网入口、部署、demo、MCP 默认地址、中国直连，先补读：
      • docs/architecture/NETWORK-TOPOLOGY.md
      • ops/nginx/*.conf
      • scripts/deploy.sh
      • scripts/deploy-demo.sh
      • scripts/deploy-edge.sh
  3. 根据脚本输出，补读最近的 issue / decision / review。
  4. 再读 Cloud memory 中最新的项目记忆。
  5. 再抽查最近 1-3 个 transcript，只看有效用户意图，不看命令噪声。
  6. 最后形成当前接手摘要。

Truth Policy

我不是新的权威副本。我只负责把新 agent 引到当前仍然活着的真相源上。

真相源优先级

遇到信息冲突时，按这个优先级判断：

  1. CLAUDE.md
  2. MEMORY.md
  3. docs/INDEX.md
  4. 最近的 docs/issues/*.md、docs/decisions/*.md、docs/reviews/**/*.md
  5. Cloud 项目目录下的 memory/
  6. transcript .jsonl

规则：

  • 不要把 skill 自己当成新的权威副本。
  • skill 负责导航，不负责复制现有知识。
  • 如果发现 source drift，必须显式指出冲突点和你最终采用的来源。

治理机制入口（ADR-030）

Towow 有一套上下文工程系统（不是 prompt），通过 PostToolUse hook 在编辑文件时自动注入相关思维框架：

  • 路由表：scripts/context_router.py — 文件路径 → 上下文片段映射
  • 片段库：scripts/context-fragments/ — 精炼的领域规则（每个 10-25 行）
  • Guard 路由：scripts/guard_router.py — 文件路径 → guard 脚本映射
  • Hook 入口：scripts/guard-feedback.py — PostToolUse 目标
  • 设计文档：docs/decisions/ADR-030-guard-signal-protocol-and-governance-reload.md

如果 PostToolUse hook 正常工作，接手者编辑文件时会自动收到该领域的上下文。如果没收到，检查 .claude/settings.json 中的 hook 配置。

公网入口 / 部署接手补充

涉及公网入口或生产部署时，先确认这些边界仍成立：

  • 对外品牌 / 全球默认 / MCP 默认入口看 <NETWORK_REDACTED>
  • 中国公开入口看 <NETWORK_REDACTED>
  • 原始 IP 只允许留在 preview、bridge、ops、部署脚本，不进入用户默认入口
  • edge 路由真相源在 ops/nginx/*.conf，不是线上手改的 /etc/nginx/conf.d/*
  • demo 发布分 prod / preview 两条通道，不能把 preview 产物当 prod 发

必读代码入口

完成文档入场后，优先进入这些路径：

  • backend/server.py
  • backend/product/protocol/
  • backend/product/matching/
  • backend/product/bridge/
  • backend/product/catalyst/
  • backend/product/openagents/

这些路径对应当前生产主链：

clarification-session -> discover -> invitation -> run -> prompt/respond -> result

transcript 噪声过滤

默认忽略这些内容：

  • local-command-caveat
  • /clear、/model 等命令操作
  • local-command-stdout
  • 纯模型切换信息
  • 非业务性的命令回显
  • 纯 Team assignment / shutdown 元信息（除非任务本身在审计 Agent Team 工作流）

你要优先提取的是：

  • 最近几次有效用户意图
  • 用户最近想推进的主线
  • 用户最近指出的线上问题
  • 用户最近改变的方向或约束

输出契约

完成接手后，先输出 4 个块，再展开细节：

  1. 当前主线
  2. 最新风险
  3. 下一步建议
  4. 缺失真相源 / source drift

默认要求：

  • 不直接复述旧知识
  • 先找最新来源
  • 先给当前状态，再展开细节
  • 发现 source drift 时必须显式指出

工具

脚本入口：

python3 .claude/skills/towow-dev-handoff/scripts/collect_handoff_context.py

可选参数：

python3 .claude/skills/towow-dev-handoff/scripts/collect_handoff_context.py \
  --cloud-root ~/.claude/projects/-Users-nature------Towow \
  --recent-sessions 3 \
  --recent-messages 3

脚本只负责收集入口，不替代你做判断。

What ships with it: 1 file

8.0 KB alongside SKILL.md, 1 of them executable

scripts/

Gives 0 of the 12 instructions most agent orchestration skills give in ~1.6k tokens

Counted across 742 of the 995 authors here whose files we hold, read 2026-08-07

  • Reference existing artifacts by path or URLin 53 of 742, across 25 files
  • Run the full test suite after integrating changesin 51 of 742, across 19 files
  • Dispatch one agent per independent problem domainin 50 of 742, across 17 files
  • Verify fixes do not conflictin 45 of 742, across 13 files
  • Include a suggested skills section in the documentin 45 of 742, across 17 files
  • Redact sensitive informationin 41 of 742, across 11 files
  • Save to the temporary directory of the operating systemin 39 of 742, across 10 files
  • Tailor the document to user-provided focus argumentsin 39 of 742, across 9 files
  • Spot check agent changes for systematic errorsin 34 of 742, across 7 files
  • Write a handoff document summarising the current conversationin 31 of 742, across 6 files
  • Assign each agent a specific scopein 23 of 742, across 8 files
  • Provide specific scope and clear goalin 23 of 742, across 5 files

Said here and by no other author read

  • run context collection script
  • read repo truth sources before derived materials
  • read recent issues and decisions from script output
  • read latest project memory in cloud
  • sample recent transcripts for valid user intent
  • produce handoff summary

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.