Single-purpose Claude Code skills orchestrating full dev lifecycle: Interview → PRD → Decomposition → TDD → Review → Archive. Inspired by Matt Pocock, OpenSpec, Spec Kit. Pairs with Graphify + Forest's CLAUDE.md.
Controlled autonomous execution loop for AppMaker backlog items. Runs only explicit autonomous slices with checklist/review gates, cost/iteration caps, and human stop points.
Close out completed feature. Verifies all backlog items done + review passed, moves backlog items to backlog/done/, moves feature folder to features/archive/<date>-<NNN-slug>/, optionally invokes retro to capture lessons. Adopts OpenSpec /opsx:archive flow. Use when all slices in feature are complete.
Deterministic cross-artifact gate for AppMaker projects. Checks PRD, decomposition, backlog, glossary, context packets, review, memory wiki, Graphify freshness, and archive readiness with PASS/FAIL/WARN output.
Compact context retrieval for codebase questions. Uses Graphify data when available, falls back to standard file discovery, and writes small context packets to appmaker/context/ for PRD/decompose/TDD/review reuse. Invoked by other commands when they need codebase awareness. User can invoke manually for ad-hoc queries.
Break PRD into vertical-slice backlog items (tracer bullets). Each slice cuts ALL layers end-to-end. Adopts Matt Pocock to-issues 1:1 plus AppMaker extensions execution_class, traces_to per AC, YAML front-matter, cycle detection, local backlog default. Use when PRD is ready and user wants to plan implementation work.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A spec + governance layer on top of Claude Code's runtime.
AppMaker doesn't replace Claude Code — it adds process discipline (PRD → AC → test traceability), deterministic gates (checklist, review compliance), and lifecycle artifacts (feature folders, archive, retro) ON TOP of what Claude Code already does well: tool calls, agent invocation, session management, context handling.
When Claude Code ships a better primitive — Agent View, /goal, /ultra-review, improved compaction — AppMaker delegates to it instead of re-implementing. The durable AppMaker value is what Claude Code doesn't do: spec discipline, AC-test traceability, constitution-enforced quality bars, glossary-canonical vocabulary, audit-trail-preserving archive flow.
Positioning (v0.2.12+):
/goal, /review + /ultra-review, compaction, plugin marketplacetraces_to AC linkage, constitution + glossary governance, retro → memory wiki, context packets, deterministic gatesOptionally pairs with Graphify for read-only codebase intelligence and Forest's andrej-karpathy-skills for universal agent baseline.
Form: Claude Code plugin at plugin/appmaker/. Skills loaded via --plugin-dir flag or marketplace install.
Convention: /appmaker:<name> (colon namespace per Claude Code plugin spec — like OpenSpec /opsx:propose).
Philosophy: minimal, single-purpose, opt-in everywhere. Delegate to Claude Code built-ins where they exist.
Status: v0.2.15. 19 skills (15 core + afk + status + token-audit + next). Per-slice review gate now part of /appmaker:next chain — no more silent slice-to-slice propagation. Test harness covers hook + glossary-extract + version SoT. See DESIGN.md.
Plugin is self-contained. No manual copy of templates or supporting files — /appmaker:init materializes everything from packaged resources.
# 1. Open project + load plugin for the session:
cd /path/to/your-project
claude --plugin-dir /Users/pawel/Projects/AppMaker/plugin/appmaker
# 2. In Claude Code session:
/help # Verify /appmaker:* commands appear in menu
/appmaker:init # Materialize appmaker/ from plugin resources
# (creates constitution, glossary, config.yaml, .appmaker-version,
# templates/, skills/tdd/, memory/wiki/, context/, checklists/,
# diagnostics/, afk/, backlog/, features/)
Re-running /appmaker:init on existing project = UPGRADE mode. Refreshes plugin-owned resources (templates, supporting files) while preserving user-owned state (constitution, glossary, memory, backlog, features, config). Detected via appmaker/.appmaker-version marker.
For permanent activation per project: add to .claude/settings.json (TBD — manual --plugin-dir per session for now).
Future: publish to plugin marketplace for /plugin install appmaker style.
/appmaker:init → bootstrap appmaker/ in fresh project
/appmaker:start "<intent>" → smart routing: feature, bug, refactor, research, review, prototype
/appmaker:grill "<topic>" → general-purpose relentless questioning (Matt's grill-me)
/appmaker:grill-brownfield → brownfield variant (reads code/docs/glossary/memory/context first)
/appmaker:interview → feature-specific entry, structured output to features/<NNN-slug>/
/appmaker:prd → synthesize PRD with Understanding section + Clarifications
/appmaker:decompose → vertical slices with execution_class, items go to backlog/
/appmaker:tdd <backlog-id> → test-first implementation per slice
/appmaker:diagnose → bug/perf diagnosis loop with repro + hypotheses + regression
/appmaker:review <scope> → invokes critic subagent (code-reviewer)
/appmaker:checklist → deterministic PASS/FAIL/WARN gate across artifacts
/appmaker:archive → close out feature, move to features/archive/
/appmaker:context "<topic>" → Graphify-aware context packet, fallback to file reads
/appmaker:feedback "<desc>" → quick capture from QA → backlog item
/appmaker:glossary → ubiquitous language (deterministic stub extraction + best-effort semantic review)
15 written.
npx claudepluginhub paweldobrzynski/appmaker --plugin appmakerDesign fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.
Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Complete developer toolkit for Claude Code