Guides adding new features, endpoints, pages, or components with a plan-before-code loop, contract clarification, and end-to-end verification.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-coding-playbook:acp-feature-addThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill when the user asks to build something new — a feature, endpoint, page, component, or capability.
Use this skill when the user asks to build something new — a feature, endpoint, page, component, or capability.
The goal is not to start typing as fast as possible.
The goal is to clarify the contract, plan the smallest viable implementation, and verify the new behavior actually works before claiming done.
Before writing code, write down:
If any of these are unclear, ask. Do not invent requirements.
For anything beyond a trivial 1-file change, enter Plan Mode (ExitPlanMode) before editing.
A good plan answers:
Get the plan approved before writing code.
For multi-step features, also create a TodoWrite list so progress is visible and each step ends in a verifiable state.
Before adding a new abstraction:
A new generic helper with one caller is a smell, not a feature.
Build the smallest version that satisfies the contract.
Do not:
If you find yourself wanting to generalize, stop and finish the concrete case first.
A new feature is not done when the file compiles.
Run, in order of preference:
If verification cannot be run in the current environment, say so explicitly.
Contract:
- ...
Plan:
- ...
Changed:
- ...
Verified:
- ...
Not verified:
- ...
Risks:
- ...
Stop and ask for confirmation if:
npx claudepluginhub bravekingzhang/agent-coding-playbook --plugin agent-coding-playbookGuides systematic implementation of new features: clarify intent with scenarios and requirements, verify assumptions via docs, analyze existing code before changes.
Orchestrates building a brand-new feature end to end — research, plan, TDD, review, and gated commit — by delegating each phase to the matching ECC agent.
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.