- Home
- Skills
- Mosif16
- Codex Skills
- Writing Plans
writing-plans_skill
- Rust
13
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill mosif16/codex-skills --skill writing-plans- SKILL.md3.4 KB
Overview
This skill creates complete, bite-sized implementation plans for engineers who have zero context about the codebase. It produces step-by-step tasks with exact file paths, full code examples, test commands, and expected outputs so an engineer can implement features safely. Plans are tailored for Rust projects and saved to docs/plans/YYYY-MM-DD-<feature-name>.md.
How this skill works
I'm using the writing-plans skill to create the implementation plan. The skill breaks a feature into tiny TDD-driven tasks (2–5 minutes each), lists which files to create or modify, supplies complete code snippets, test commands with expected results, and exact git commit commands. It assumes the implementer knows programming but not the toolset or domain; instructions avoid assumptions and include verification steps and doc locations.
When to use it
- Design is finished and you need executable implementation steps for engineers with no repo context
- Hiring contractors or rotating team members into a feature quickly
- Preparing a handoff to an engineer unfamiliar with Rust or the project
- When you want strict TDD, small commits, and reproducible verification
- Creating audit-ready development plans for critical features
Best practices
- Start the plan with the required header and save to docs/plans/YYYY-MM-DD-<feature-name>.md
- Make each step one atomic action: write failing test, run, implement minimal code, run, commit
- Always include exact file paths (create/modify/test) and full code examples — no placeholders
- Prefer minimal implementations (YAGNI) and repeatable test commands with expected output
- DRY task descriptions and reference required sub-skills (e.g., superpowers:executing-plans) when handing off execution
Example use cases
- Add a new serialization layer: plan includes src/serial/mod.rs, tests/serial/test_basic.rs, and cargo test commands with expected output
- Implement feature flag evaluation: tasks create src/flags.rs, unit tests, example config file, and stepwise commits
- Migrate a struct to serde derive: specify exact file edits, full before/after code, tests to run, and git history steps
- Introduce an API endpoint in a Rust service: list handler file, integration test, cargo run instructions, and verification curl commands
FAQ
Plans are saved to docs/plans/YYYY-MM-DD-<feature-name>.md using the exact filename you declare at plan creation.
Must I follow TDD order?
Yes. Each task must be written as a failing test first, then minimal implementation, then verification and commit.
How granular should tasks be?
Each task should be a single action that takes ~2–5 minutes: write one test, run it, implement minimal code, run tests, commit.