跳到正文
cccskills
免费在 GitHub 公开

devprocess

Structured build workflow: branch, delegate to coding agent, test, screenshot, and commit. Triggers on: "$devprocess", "build this properly", "use the coding agent workflow", or any feature/refactor/infrastructure change that is too large for a direct edit. NOT for: one-line fixes, doc edits, prompt tweaks, or simple config changes — do those directly.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md3.7 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

$devprocess — How We Build

Core Principle

You are the product partner. The coding agent is the worker for substantial code changes. Do the framing, architecture, review, verification, and release decisions here. Delegate the heavy implementation work.

Small direct edits are fine for docs, prompts, tiny fixes, or surgical config changes. For feature work, refactors, or anything that needs exploration, use a coding agent.

Before Every Feature

1. Read the Code First

  • Read the repo structure and the relevant files before deciding anything.
  • Reuse existing patterns before inventing new ones.
  • For architecture questions, read graphify-out/GRAPH_REPORT.md first if it exists.

2. Create a Branch

git checkout -b <prefix>/<feature-name>
  • Do not build features directly on main.
  • One branch per logical change.

3. Write the Plan

Create plans/<branch-name>/plan.md with:

  • What — what is being built
  • Why — what problem it solves
  • How — approach, touched files, constraints
  • Tests — what must pass
  • Definition of done — what proves completion

If the request is ambiguous or long-running, also create or update a lightweight spec first.

During Development

4. Delegate the Build

  • Use the coding-agent skill for substantial implementation.
  • Give the coding agent the plan, target files, constraints, and definition of done.
  • If the first pass is poor, tighten the brief and rerun.
  • Do not accept unexplained large diffs.

5. Test

  • Add or update tests with the change when appropriate.
  • Run the smallest meaningful gate first, then the broader suite if needed.
  • Do not claim success if tests are red.

6. Look at the Real UI

  • Open the app or rendered output when there is any user-facing change.
  • Take screenshots or snapshots when that is the best proof.
  • Browser-visible bugs do not count as fixed until inspected visually.

7. Commit in Small Units

  • One logical change per commit.
  • Review the diff before each commit.
  • Push often enough that work is safe.

After Building

8. Measure

Capture before/after evidence when possible:

  • tests added/passing
  • load time or latency change
  • bug reproduced then gone
  • screenshot before/after
  • error rate change

9. Review the Diff Yourself

  • Read the patch like a reviewer.
  • Check naming, duplication, edge cases, and fit with existing patterns.
  • Revert sloppy code and rerun if needed.

10. Merge or PR

  • If the change is verified and low risk: merge and ship.
  • If risk or uncertainty remains: open a PR with a concise summary and ask for review.

OpenClaw-Specific Defaults

  • Prefer sessions_spawn with runtime="acp" when the owner explicitly wants Codex/Claude Code/Cursor/Gemini-style harness work.
  • Prefer the bundled coding-agent workflow for substantial repo work when local execution is the right fit.
  • After changing code files, run graphify update . when this repo uses graphify.
  • Always leave the branch, plan, commits, and verification artifacts in a state another person can inspect.

Anti-Patterns

  • Writing a whole feature yourself without first deciding whether it should be delegated
  • Starting on main
  • No plan file
  • One giant unreviewed commit
  • Saying "tests pass" without running them
  • Saying UI is good without opening it
  • Shipping local-only work and calling it done

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

ai-pa

免费

AI Personal Assistant network skill for multi-agent PA coordination. Use when: contacting another PA, coordinating with peer agents, scheduling meetings between owners, broadcasting messages to PA groups, or looking up contacts from the local PA directory. Reads contact data from data/pa-directory.json in the workspace.

中文简介正在准备中。当前显示原文说明。

netanel-abergel/pa-skills62026年5月10日 更新

ai-pa

免费

AI Personal Assistant network skill. Use when: contacting another PA, coordinating with peer agents, scheduling meetings between owners, broadcasting messages, looking up contacts from the local PA directory, or performing any cross-PA task. Reads contact data from data/pa-directory.json in the workspace. Does NOT contain hardcoded contact data — each PA maintains their own directory file.

中文简介正在准备中。当前显示原文说明。

netanel-abergel/pa-skills62026年5月10日 更新

ai-pa

免费

Route and execute inter-PA coordination tasks in the OpenClaw multi-agent PA network. Triggers when: contacting another PA, passing a message between owners, broadcasting to a PA group, handling group lead/silence rules, or adding a new PA to the directory. NOT for: contact lookups (use contact-list skill), calendar invites (use meetings skill), or billing issues (use billing-monitor skill).

中文简介正在准备中。当前显示原文说明。

netanel-abergel/pa-skills62026年5月10日 更新

OpenClaw API gateway proxy for authenticated SaaS calls (100+ services in references/SERVICES.md). Use ONLY when making an HTTP request to a third-party service that requires the gateway's stored credentials. NOT for direct API calls with bearer tokens already in scope, NOT for monday.com (use monday-for-agents), NOT for WhatsApp (use heleni-whatsapp). Triggers: "call the X API", "fetch from <service>", "send to Slack via gateway", "use the gateway for".

中文简介正在准备中。当前显示原文说明。

netanel-abergel/pa-skills62026年5月10日 更新

Automatically creates new skills from complex problem-solving sessions. Triggers at the end of multi-step tasks that required novel approaches, debugging sessions, or workflow creation. Use when: a task took 3+ tool calls, involved a non-obvious solution, or established a pattern likely to recur. Also triggers on: "create a skill from this", "save this as a skill", "this should be a skill".

中文简介正在准备中。当前显示原文说明。

netanel-abergel/pa-skills62026年5月10日 更新

Monitor for API billing errors and alert the owner and admin immediately. Use when: an API billing error is detected, a peer PA reports a billing error, or during routine health checks. Handles detection, notification, and fallback model switching. Model-agnostic: works with any LLM provider (Anthropic, OpenAI, Google, etc.).

中文简介正在准备中。当前显示原文说明。

netanel-abergel/pa-skills62026年5月10日 更新

查看 netanel-abergel 的全部技能

报告此技能的问题