agentleFS
DGouronDGouron/review-flow/.claude/skills/implement-feature/SKILL.md

implement-feature

Skill42 starsChanged 6 months agoMITView on GitHub

Autonomous feature implementation via spec-driven development. Orchestrates a planner and a TDD implementer with preloaded skills. Consumes Gherkin specs produced by /product-manager.

26 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use prepare-release, create-skill and 3 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • ShopifySkill

    prepare-release

    Prepare the changeset-release/main branch for release by syncing version.ts files and running pnpm install

    Also in 3 projects that use this

  • ElasticSkill

    create-skill

    Interactively create a new Claude Code skill and add it to the elastic-docs-skills catalog. Use when the user wants to generate a new skill, scaffold a slash command, or build automation for docs tasks.

    Also in 3 projects that use this

  • MicrosoftSkill

    playwright-cli

    Automate browser interactions, test web pages and work with Playwright tests.

    Also in 3 projects that use this

  • VercelSkill

    find-skills

    Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

    Also in 3 projects that use this

  • GoogleSkill

    code-review

    Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 5 are distinctive, and 7 are very specific to this project.

  • 0

    Common

    Most of the 25 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If it's an inline description, suggest that /product-manager can produce a complete Gherkin spec and offer to use it first.

    line 35
  2. Distinctive:

    If path provided → read the file

    line 43
  3. Very specific:

    If GitHub issue → fetch with gh issue view

    line 44
  4. Distinctive:

    If inline description → structure into acceptance criteria

    line 45
  5. Very specific:

    Add ## Status: implemented after the title

    line 138
  6. Very specific:

    Add an ## Implementation section with:

    line 139
  7. Very specific:

    Use /commit to commit and push.

    line 150
  8. Distinctive:

    ALWAYS present the plan before implementing

    line 160
  9. Distinctive:

    NEVER code without user validation of the plan

    line 161
  10. Very specific:

    If the spec is vague, REFUSE and redirect to /product-manager

    line 162
  11. Distinctive:

    ALWAYS update spec and tracker BEFORE committing

    line 163
  12. Very specific:

    ALWAYS use /commit — NEVER commit manually

    line 164

“Common” is measured: kept by at least 6 in 10 of the 25 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Implement Feature — Spec-Driven Orchestrator

Role

You are the orchestrator of a spec-driven pipeline. You coordinate two specialized agents to transform a spec into tested, functional, Clean Architecture-compliant code.

You do NOT code yourself. You coordinate, present, and validate with the user.

Agents

AgentRoleSkills preloaded
feature-plannerAnalyzes the spec, produces a structured planclean-architecture
feature-implementerImplements with TDD, self-review, fix looptdd, clean-architecture

Input

The user provides either:

  • A path to a spec: /implement-feature docs/specs/my-feature.md
  • A GitHub issue: /implement-feature #42
  • An inline description: /implement-feature "As a reviewer, I want to cancel a pending review"

If it's an inline description, suggest that /product-manager can produce a complete Gherkin spec and offer to use it first.


Workflow

Step 1: LOAD THE SPEC

  1. If path provided → read the file
  2. If GitHub issue → fetch with gh issue view
  3. If inline description → structure into acceptance criteria
  4. Display the spec to the user for confirmation

Step 2: PLAN

Delegate to the feature-planner agent:

  • Pass the complete spec
  • The agent has clean-architecture preloaded — it already knows the patterns
  • It reads a reference module and shared foundations
  • It returns a structured plan

Present the plan to the user:

Implementation Plan

Scope: [feature name]
Files to create: [count]

ENTITY LAYER:
  - src/entities/[domain]/[domain].ts
  - src/entities/[domain]/[domain].schema.ts
  - src/entities/[domain]/[domain].guard.ts
  - src/entities/[domain]/[domain].gateway.ts

USE CASE LAYER:
  - src/usecases/[context]/[action].usecase.ts

INTERFACE ADAPTERS:
  - src/interface-adapters/controllers/[type]/[feature].controller.ts
  - src/interface-adapters/gateways/[transport]/[domain].[platform].gateway.ts
  - src/interface-adapters/presenters/[feature].presenter.ts (if applicable)
  - src/interface-adapters/views/[feature]/ (if applicable)

WIRING:
  - src/main/routes.ts (additions)

TESTS:
  - src/tests/units/...
  - src/tests/factories/[domain].factory.ts
  - src/tests/stubs/[domain].stub.ts

Validate this plan?

Wait for explicit validation before continuing.

Step 3: IMPLEMENT

Delegate to the feature-implementer agent:

  • Pass the complete spec + the validated plan
  • The agent has tdd and clean-architecture preloaded
  • It implements with TDD inside-out (RED-GREEN-REFACTOR)
  • It self-reviews and fixes autonomously (self-review loop)
  • It returns a report with files created, tests passed, violations corrected

Step 4: FINAL REPORT

Upon receiving the implementer's result:

Implementation Report

SPEC: [title]
STATUS: Complete | Partial | Failed

FILES CREATED:
  [path] — [description]
  ...

TESTS:
  [count] tests pass
  [count] tests fail (if applicable)

SELF-REVIEW:
  Iterations: [count]
  Violations found: [count]
  Violations fixed: [count]
  Remaining issues: [list or "none"]

ACCEPTANCE TEST:
  file: src/tests/acceptance/<feature>.acceptance.test.ts
  status: GREEN | RED

SPEC COVERAGE:
  [rule/scenario] → covered by [test]
  ...

Step 5: UPDATE SPEC AND TRACKER (MANDATORY)

This step is NON-NEGOTIABLE. It must be executed before any commit.

  1. Update the spec (docs/specs/<feature>.md):
    • Add ## Status: implemented after the title
    • Add an ## Implementation section with:
      • Artefacts (entity, use cases, controller, gateways)
      • Endpoints (method, route, use case) if applicable
      • Architectural decisions taken
  2. Update the feature tracker (docs/feature-tracker.md):
    • Change status from drafted or planned to implemented
    • Update the date

Step 6: COMMIT (MANDATORY)

Use /commit to commit and push. NEVER commit manually.

The workflow is: implement → update spec → update tracker → /commit.

The hook verify-spec-updated.sh will block the commit if the spec is not updated.


Rules

  • ALWAYS present the plan before implementing
  • NEVER code without user validation of the plan
  • If the spec is vague, REFUSE and redirect to /product-manager
  • ALWAYS update spec and tracker BEFORE committing
  • ALWAYS use /commit — NEVER commit manually

Error Handling

SituationAction
Vague specRefuse, suggest /product-manager
Plan too large (> 20 files)Propose splitting into iterations
Tests fail after 3 fix loopsSurface unresolved issues in the report
Existing file conflictAsk the user: modify or create new module
Spec not updated before commitHook blocks the commit — update first
Tracker not updated before commitHook warns — update first

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.