waterfall-plan
Skill0 starsChanged 53 days agoMITView on GitHub
Detail a chosen design approach into an implementation plan (objective, file manifest, execution order, test strategy), then auto-proceed to waterfall-challenge. Convergent phase of the waterfall flow.
In this repository26 other files it gives its agents.
- ask-mattSkill
- code-reviewSkill
- create-prSkill
- draft-prSkill
- ready-prSkill
- developSkill
- self-review-docsSkill
- implement-reviewSkill
- implementSkill
- implement-specSkill
- librarianSkill
- nemesisSkill
- refactorSkill
- research-codebaseSkill
- researchSkill
- review-docsSkill
- self-reviewSkill
- simplifySkill
- tddSkill
- test-auditSkill
- waterfall-challengeSkill
- waterfall-designSkill
- waterfall-implementSkill
- waterfall-refactorSkill
- waterfall-reviewSkill
- zero-tech-debtSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 45
If detailing the approach reveals problems not visible at the design level (e.g.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Plan
Take the approach chosen in the waterfall-design skill and produce a
detailed implementation plan.
This is the convergent phase. The option space has already been exhausted in design; do not reopen it here. Detail the chosen approach into files, execution order, and tests. If detailing reveals the approach is wrong, surface it (Step 2) rather than silently switching.
Prerequisites
This skill expects an approach to have been selected via the
waterfall-design skill. If no chosen approach exists in the
conversation, ask the user what to plan.
Step 1: Extract the chosen approach
Pull the selected approach from the conversation, including:
- What the approach does and why it was chosen.
- Which files are affected.
- Known trade-offs and risks.
Step 2: Detail the plan
Expand the approach into a concrete implementation plan:
- Objective: One sentence stating what the feature does.
- Approach: How it will be built, at the level of "which modules change and why."
- File manifest: Files to create, modify, or remove, with a one-line note per file describing what changes.
- Test strategy: What tests to write and at which level (unit, integration, e2e).
- Open questions: Anything unresolved that the challenge phase should examine.
If detailing the approach reveals problems not visible at the design level (e.g. a dependency that makes the approach impractical), note them clearly. Do not silently switch to a different approach — surface the issue.
Step 3: Slice into increments
Structure the plan into shippable increments. Each increment:
- Is a single commit that leaves the codebase in a working state.
- Has a clear purpose describable in a commit message.
- Is independently reviewable — a reviewer can understand it without reading the other increments.
- Includes the relevant tests for the behavior it introduces.
Order the increments so each one builds on the last. Prefer:
- Refactoring pre-work first (from the design phase's refactor analysis).
- Infrastructure and types before the code that uses them.
- Core logic before convenience features or edge case handling.
- Tests alongside the code they cover, not in a separate increment.
Present the increments as a numbered list. For each, note which files change and what the commit message would be. If the whole feature is small enough for a single commit, say so — do not split for the sake of splitting.
Step 4: Present and proceed to challenge
Present the plan to the user, then invoke the waterfall-challenge skill
to pressure-test it. Do not wait for approval — the challenge phase is
what validates the plan. The user will approve (or reject) after the
challenge.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

