writing-plans
About
The writing-plans skill creates detailed implementation plans for engineers after design completion, particularly when they lack codebase context. It generates comprehensive tasks with exact file paths, complete code examples, and verification steps. This skill assumes minimal domain knowledge and emphasizes TDD, frequent commits, and bite-sized implementation tasks.
Documentation
Writing Plans
Overview
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
Zero-Context Test
Before finalizing ANY plan, verify:
Can someone execute this if they:
□ Never saw our codebase
□ Don't know our framework
□ Only have this document
□ Have no context about our domain
If NO to any → Add more detail
Every task must be executable in isolation.
Bite-Sized Task Granularity
Each step is one action (2-5 minutes):
- "Write the failing test" - step
- "Run it to make sure it fails" - step
- "Implement the minimal code to make the test pass" - step
- "Run the tests and make sure they pass" - step
- "Commit" - step
Plan Document Header
Every plan MUST start with this header:
# [Feature Name] Implementation Plan
> **For Claude:** REQUIRED SUB-SKILL: Use ring: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]
**Global Prerequisites:**
- Environment: [OS, runtime versions]
- Tools: [Exact commands to verify: `python --version`, `npm --version`]
- Access: [Any API keys, services that must be running]
- State: [Branch to work from, any required setup]
**Verification before starting:**
```bash
# Run ALL these commands and verify output:
python --version # Expected: Python 3.8+
npm --version # Expected: 7.0+
git status # Expected: clean working tree
pytest --version # Expected: 7.0+
## Task Structure
```markdown
### 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`
**Prerequisites:**
- Tools: pytest v7.0+, Python 3.8+
- Files must exist: `src/config.py`, `tests/conftest.py`
- Environment: `TESTING=true` must be set
**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 output:
FAILED tests/path/test.py::test_name - NameError: name 'function' is not defined
If you see different error: Check file paths and imports
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 output:
PASSED tests/path/test.py::test_name
Step 5: Commit
git add tests/path/test.py src/path/file.py
git commit -m "feat: add specific feature"
**If Task Fails:**
1. **Test won't run:**
- Check: `ls tests/path/` (file exists?)
- Fix: Create missing directories first
- Rollback: `git checkout -- .`
2. **Implementation breaks other tests:**
- Run: `pytest` (check what broke)
- Rollback: `git reset --hard HEAD`
- Revisit: Design may conflict with existing code
3. **Can't recover:**
- Document: What failed and why
- Stop: Return to human partner
- Don't: Try to fix without understanding
## 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
## Required Patterns
This skill uses these universal patterns:
- **State Tracking:** See `skills/shared-patterns/state-tracking.md`
- **Failure Recovery:** See `skills/shared-patterns/failure-recovery.md`
- **Exit Criteria:** See `skills/shared-patterns/exit-criteria.md`
- **TodoWrite:** See `skills/shared-patterns/todowrite-integration.md`
Apply ALL patterns when using this skill.
## Execution Handoff
After saving the plan, offer execution choice:
**"Plan complete and saved to `docs/plans/<filename>.md`. Two execution options:**
**1. Subagent-Driven (this session)** - I dispatch fresh subagent per task, review between tasks, fast iteration
**2. Parallel Session (separate)** - Open new session with executing-plans, batch execution with checkpoints
**Which approach?"**
**If Subagent-Driven chosen:**
- **REQUIRED SUB-SKILL:** Use ring:subagent-driven-development
- Stay in this session
- Fresh subagent per task + code review
**If Parallel Session chosen:**
- Guide them to open new session in worktree
- **REQUIRED SUB-SKILL:** New session uses ring:executing-plans
Quick Install
/plugin add https://github.com/LerianStudio/ring/tree/main/writing-plansCopy and paste this command in Claude Code to install this skill
GitHub 仓库
Related Skills
llamaguard
OtherLlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.
sglang
MetaSGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.
evaluating-llms-harness
TestingThis Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.
langchain
MetaLangChain is a framework for building LLM applications using agents, chains, and RAG pipelines. It supports multiple LLM providers, offers 500+ integrations, and includes features like tool calling and memory management. Use it for rapid prototyping and deploying production systems like chatbots, autonomous agents, and question-answering services.
