用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/kitfunso/claude-config --skill writing-plans命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
| name | writing-plans |
| description | Use when you have a spec or requirements for a multi-step task, before touching code |
Write comprehensive implementation plans assuming the engineer has zero context for our codebase and questionable taste. Document everything they need to know: which files to touch for each task, code, testing, docs they might need to check, how to test it. Give them the whole plan as bite-sized tasks. DRY. YAGNI. TDD. Frequent commits.
Assume they are a skilled developer, but know almost nothing about our toolset or problem domain. Assume they don't know good test design very well.
Announce at start: "I'm using the writing-plans skill to create the implementation plan."
Context: This should be run in a dedicated worktree (created by brainstorming skill).
Save plans to: docs/plans/YYYY-MM-DD-<feature-name>.md
Plans that enumerate specific call sites or other concrete code targets must READ the code before constructing the per-site mapping. Reasoning from names and line proximity is reliably wrong.
Before listing per-site fixes, grep the codebase for:
grep -r <obvious-name> first — resolveTenantId, requireAuth, withDb, etc. — saves a plan-stage retry.cmdX may call api.x(ctx, id) which calls the primitive; the per-site mapping must target the primitive's actual direct caller, not the surface command the user types.function <name> and bound the ranges. Mislabels make execute rediscover the enclosing function instead of following the plan.A plan that skips these is reliably caught at the plan-eng-critic gate and triggers a retry. The grep cost is far below the retry cost. Learned 2026-05-23 from the hippo tenant-isolation residue plan Revision 0, which made all four of these mistakes and failed plan-eng at score 58.
Each step is one action (2-5 minutes):
Every plan MUST start with this header:
# [Feature Name] Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use executing-plans to implement this plan task-by-task.
**Goal:** [One sentence describing what this builds]
**Architecture:** [2-3 sentences about approach]
**Tech Stack:** [Key technologies/libraries]
---
### Task N: [Component Name]
**Files:**
- Create: `exact/path/to/file.py`
- Modify: `exact/path/to/existing.py:123-145`
- Test: `tests/exact/path/to/test.py`
**Step 1: Write the failing test**
```python
def test_specific_behavior():
result = function(input)
assert result == expected
Step 2: Run test to verify it fails
Run: pytest tests/path/test.py::test_name -v
Expected: FAIL with "function not defined"
Step 3: Write minimal implementation
def function(input):
return expected
Step 4: Run test to verify it passes
Run: pytest tests/path/test.py::test_name -v
Expected: PASS
Step 5: Commit
git add tests/path/test.py src/path/file.py
git commit -m "feat: add specific feature"
## Remember
- Exact file paths always
- Complete code in plan (not "add validation")
- Exact commands with expected output
- Reference relevant skills with @ syntax
- DRY, YAGNI, TDD, frequent commits
## Execution Handoff
After saving the plan, announce: "Plan complete and saved to `docs/plans/<filename>.md`."
Then pick the execution mode yourself — auto-decide, no menu:
- **Default — Subagent-Driven (this session):** fresh subagent per task, review between tasks. REQUIRED SUB-SKILL: subagent-driven-development.
- **Parallel Session** only when the plan is too large for this session's remaining context, or the user asked for a separate session: guide them to open a new session in the worktree. REQUIRED SUB-SKILL: executing-plans.
Name the choice in one line and start. The user can redirect.