claudecodex

implementation-workflow

A multi-agent implementation pipeline that takes a GitHub Map Issue task (or a standalone requirement) through user scrutiny, optional Issue refinement, codebase investigation, planning, a human-approved test-first specification (automated tests plus committed manual test procedures), implementation against frozen tests, automated and human review, history cleanup, and PR.

0 on GitHubby dangov3.5.0MITupdated 13 days agoProductivity
/

Plugin implementation-workflow from marketplace k1nak0 in k1nak0/marketplace. Commands read from the repo's manifest — 12 days ago.

What it installs

Skills (5)

  • implementation-planning

    Synthesise requirements, codebase analysis, and library research into a concrete implementation plan that splits every acceptance criterion into automated test cases and manual test steps, plus a CI readiness check and a documentation update plan. Runs inline (not as an isolated agent) so it can ask

  • implementation-workflow

    Main entry point for implementation-workflow. Takes a task off a GitHub Map Issue (or a standalone requirement) through user scrutiny, optional Issue refinement, codebase investigation, planning, human-approved test-first specification, implementation against frozen tests, automated and human review

  • issue-refinement

    Runs when the user rejects a task's content at the Phase 1 scrutiny gate. Talks the change through with the user, then rewrites the Task Issue, updates every Issue and Map Issue row the change affects, and brings the docs/design/ documents back in line — shipping the doc changes as their own PR. Use

  • onboarding

    Set up a project to use implementation-workflow — verifies gh CLI/GitHub remote prerequisites, reports CI and docs/adr readiness, checks for a root CLAUDE.md, and creates, updates, or migrates the project's docs/tools/ directory. Run once when adopting the plugin in a new project, or whenever projec

  • requirement-understanding

    Establish what to implement this run — let the user choose a task from a GitHub Map Issue (or interview them for a standalone requirement), then put the task's content in front of them for scrutiny before any work starts. Returns a verdict of ready or needs-refinement. Writes requirements-report.md.

Subagents (7)

  • code-reviewer

    Performs a fresh, critical review of all changed files against the frozen tests, the implementation plan, and CLAUDE.md. First verifies mechanically that neither the frozen test files nor the frozen manual-test documents were modified — any change is Critical. Also checks that the why was recorded i

  • implementer

    Makes the frozen, human-approved automated tests pass and executes the frozen manual tests. Must never modify test files or manual-test documents, weaken assertions, or change CI so a frozen test stops running; escalates with a test-dispute.md instead. Never runs git. Records the why for each decisi

  • library-researcher

    Retrieves up-to-date library documentation using WebSearch/WebFetch (or a project-declared docs MCP) to identify correct API patterns and usage examples for a required external library. Produces a library-usage-report.md. Use for Phase 5 (Library Investigation) only when requirements-report.md indic

  • persistence-engineer

    Commits the implementer's uncommitted work as the canonical test+implementation commit series, flips this run's draft ADRs to accepted, pushes (force-with-lease only after a rewrite of an already-pushed branch), and opens or updates the Pull Request via gh. Puts the how-narrative in the PR body, nev

  • repository-explorer

    Investigates a codebase using Grep/Glob/Read to identify files, symbols, and reference chains relevant to a feature request. Produces an impact-analysis-report.md. Use for Phase 4 (Codebase Investigation) after Phase 3 has cut the work branch.

  • test-reviewer

    Performs a fresh, critical review of the frozen-test candidate a test-writer just produced — before any human sees it. Checks that automated tests assert observable behaviour rather than internals, that every acceptance criterion is covered by the right kind of test, that manual-test steps have conc

  • test-writer

    Writes the specification of a task's behaviour before any implementation exists — automated tests for what a runner can check, and a committed manual-test document under docs/manual-tests/ for what it can't. Never writes production code beyond signature-only scaffolding. Adds CI configuration if the

View source on GitHub

More from k1nak0/marketplace

All 3 plugins →

task-splitter

claudecodex

Splits an epic into PR-sized tasks and registers them as a GitHub Map Issue plus per-task Issues. In design mode it interviews the user, writes a behavior-only design doc with an ADR for each decision that doc states as a fact, and ships those docs as their own PR. In split mode it reads a design doc that already exists and goes straight to the breakdown.

5 skills

0

light-workflow

claudecodex

A single-skill lightweight implementation workflow: branch preparation, a requirement discussion with the user, implementation that consults the user whenever it hits a real fork, one human approval gate, then commit, push, and PR. No test freeze, no automated review gate, no Issue registration — but the same VCS minimalism: source code is the only *how*, and every *why* lands in a comment, a commit message, or an ADR.

1 skill

0

Similar plugins

Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).

39 skills

mattpocock/skills 323,108

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

caveman

claude

Talk like caveman. Cut filler. Keep technical details explicit.

22 skills · 1 command · 3 agents · hooks

JuliusBrussee/caveman 119,462

claude-mem

claudecodex

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

Taste Skill

claudecodex

tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.

1 skill

Leonxlnx/taste-skill 114,391

career-ops

claudecodex

Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications

5 skills · hooks

career-ops-hq/career-ops 82,253

Maintain implementation-workflow?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins3 plugins