feature-dev

Implements full-lifecycle feature development from requirements analysis through verified git commits.

1|Updated Sep 11, 2026
One-click install
npx skills add https://github.com/lldwb/lldwb-claude-skills --skill feature-dev-lldwb
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: feature-dev
Source: https://github.com/lldwb/lldwb-claude-skills/tree/main/skills/feature-dev
Command: npx skills add https://github.com/lldwb/lldwb-claude-skills --skill feature-dev-lldwb

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Turning a feature request into verified, committed code involves many failure-prone steps: misunderstood requirements, unverified assumptions about third-party dependencies, missing plans, and untested integrations. This Skill enforces a disciplined end-to-end workflow so features are analyzed, designed, implemented, tested, and committed with evidence at every stage. ## Core Features & Use Cases - Structured Requirements Analysis: Scores requirement completeness across four dimensions (goal clarity, expected outcome, scope, constraints) and asks targeted clarifying questions before design begins. - Feasibility Verification: Verifies third-party component capabilities, configuration keys, and protocol behaviors using read-only evidence (decompilation, reflection, minimal reproduction) instead of assumptions. - Planning Documents: Produces a proposal/design/tasks three-document set with checkbox-tracked tasks when the change spans modules or requires multiple iterations. - Layered Implementation & E2E Verification: Implements bottom-up (model → service → controller → frontend) with compile, unit test, and browser-MCP end-to-end verification, then commits via conventional git commits. - Use Case: A user asks to add a new API endpoint spanning backend and frontend. The Skill analyzes the requirement, verifies external dependencies, drafts a plan for confirmation, implements layer by layer, validates through the real browser link, and commits locally. ## Quick Start Ask the agent to take this new feature requirement from analysis through implementation, end-to-end verification, and a local git commit using the feature-dev workflow.

Frequently Asked Questions about feature-dev

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I manage a full feature development workflow with an AI agent?

Use a staged workflow: analyze and score the requirement, verify external dependency feasibility with read-only evidence, present a design for user confirmation, implement layer by layer, then verify with compile, unit tests, and end-to-end browser testing before committing.

How to verify third-party library capabilities before writing code?

Inspect the actual artifacts instead of assuming: use unzip and javap to read class signatures and bytecode, reflection to list methods, and check existing callers in the project. Confirm configuration keys against real @Value annotations rather than guessing.

When should feature development produce planning documents?

Write the proposal/design/tasks document set when changes span modules or frontend and backend, involve external dependencies, need multiple iterations, or require checklist-style acceptance. Single-file, one-round changes can stay as an in-conversation plan.

Does this workflow support interactive stage-by-stage confirmation?

Yes, an optional six-phase mode (research, ideation, planning, execution, optimization, review) pauses for user confirmation at each stage. The review phase dispatches a read-only subagent to independently verify completion against acceptance criteria.

When should I not use a full feature development workflow?

Skip it for small-scope code optimizations, bug fixes, comment-only changes, documentation-only updates, or pure frontend error diagnosis. Those scenarios route to lighter specialized workflows instead of the full requirements-to-commit pipeline.