incremental-implementation
Delivers changes incrementally in thin, verifiable slices
by addyosmaniFree
Description
incremental-implementation is an agent skill from Addy Osmani's agent-skills collection for delivering a change in thin, verifiable slices. The method is to build one piece, test it, verify it and then expand, rather than implementing a whole feature in one pass. Each increment should leave the system in a working, testable state. The skill calls this the execution discipline that makes large features manageable. It applies to any multi-file change, to building a feature from a task breakdown, to refactoring existing code, and any time the agent is tempted to write more than about a hundred lines before testing. It says not to use it for single-file, single-function changes where the scope is already minimal. The skill describes an increment cycle that the agent repeats for each slice. It is listed on skills.sh, with its security audit results shown there. It suits developers and agents who implement features across several files and want each step to be safe to stop after, so a failure is easy to locate and undo.…
Highlights
- Thin vertical slices of work
- Each increment leaves a working system
- Increment cycle: build, test, verify, expand
- Guidance on when not to use it
Ratings & Reviews
No ratings yet
Be the first to rate incremental-implementation.
People also visit
migrate-to-shoehorn
Migrate test files from `as` type assertions to @total-typescript/shoehorn
benchmark-optimization-loop
Convert 'make it faster' requests into a bounded measured optimization loop
cpp-testing
Use only when writing/updating/fixing C++ tests, configuring GoogleTest/CTest
incremental-implementation FAQ
What does the incremental-implementation skill do?
It guides an agent to deliver a feature or change in thin, verifiable slices, each one tested before the next starts.
How do I install it?
Run: npx skills add https://github.com/addyosmani/agent-skills --skill incremental-implementation
When should I use it?
For multi-file changes, features built from a task list, refactors, and any time the agent is about to write more than about a hundred lines untested.
When should I skip it?
For single-file or single-function changes where the scope is already minimal.
What security checks has it passed?
skills.sh lists three audits: Gen Agent Trust Hub Pass, Socket Pass and Snyk Pass.
What are the best incremental-implementation alternatives?
Popular alternatives to incremental-implementation include migrate-to-shoehorn, tdd and ai-regression-testing. See all incremental-implementation alternatives