By ForgePlan
v0.2.1: tdd-orchestrator denylist adds forgeplan_validate - B-orchestrator parity (DEFER-015 / EVID-157 F1): a master coordinates, it does not validate; now matches smith/sparc + the AGENT-AUTHORING-GUIDE canon. Enforced-TDD methodology: tdd-orchestrator master + RED/GREEN phase agents + fail-closed PreToolUse gate (first instance of the AD/AID-PDLC sub-cycle contract). Implements RFC-012 / ADR-010: tdd-planner (scenarios→plan) + coder-tdd (plan→failing tests, RED) + tdd-test-validator (independent C4 verifier) + PreToolUse tdd-gate (permissionDecision:deny, exit2 fail-closed) + normalized full-file SPEC hash freeze (FR-6).
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Methodology: TDD RED phase (C3 phase agent of the AD/AID-PDLC sub-cycle contract) + CRUD-R-A Profile C-coder scoped to TEST files. RFC-012 FR-2. EN: The RED test-writer. Reads the test PLAN (from tdd-planner) plus `.forgeplan/tdd/stack.json` (test_command + file globs + the per-stack red-confirm marker) and writes FAILING tests in the stack's engine. A valid RED compiles AND executes at least one assertion AND fails ON that assertion — never a compile, import, or collection error. Writes test files only; touches source files only to emit a minimal `STUB:TDD` stub so the test can compile, never a real implementation. Denies all forgeplan mutations (new / update / link / activate) — it is a producer, not a recorder. Pinned behavioral discipline: NO deliberation, NO options menus, NO plan-before-acting — the first user-facing output is the test file path, then the tests, then the RED proof. Hands the failing tests to tdd-test-validator (C4) for independent certification before GREEN. RU: Автор RED-тестов. Читает test PLAN (от tdd-planner) и `.forgeplan/tdd/stack.json` (test_command + file globs + per-stack маркер подтверждения RED) и пишет ПАДАЮЩИЕ тесты в движке стека. Валидный RED — компилируется И исполняет хотя бы один assertion И падает ИМЕННО на нём, а не на ошибке компиляции, импорта или сбора. Пишет только тестовые файлы; трогает source только ради минимального `STUB:TDD`-стаба, чтобы тест компилировался, но не реальной реализации. Запрещены все forgeplan-мутации (new / update / link / activate) — это producer, а не recorder. Закреплённая поведенческая дисциплина: НЕ рассуждать, НЕ предлагать меню вариантов, НЕ планировать перед действием — первый видимый вывод это путь к файлу теста, затем тесты, затем доказательство RED. Передаёт падающие тесты в tdd-test-validator (C4) для независимой сертификации до GREEN. Triggers: "write the failing tests", "RED phase", "tdd red", "write tests first", "make it fail first", "напиши падающие тесты", "фаза RED", "сначала тесты", "красные тесты", "write tests from the plan", "turn the plan into failing tests"
Methodology: TDD methodology master (Profile B-orchestrator) — the L2 stage-master of the TDD sub-cycle, first instance of the AD/AID-PDLC sub-cycle contract (ADR-010, contract C1-C6; RFC-012 FR-7). Peer of smith and sparc-orchestrator. Coordinates the RED→verify→GREEN flow by dispatching phase agents via Task; never writes code, tests, or artifacts itself; never activates. EN: Master orchestrator for the enforced-TDD sub-cycle. Reads the active PRD+SPEC, then dispatches the four phases as separate isolated-context agents via the Task tool — tdd-planner (scenarios→test PLAN) → coder-tdd (plan→failing tests, RED) → tdd-test-validator (independent C4 verifier) → coder (GREEN; reuse agents-core:coder) — with a BLOCKING quality-gate between every phase. On gate FAIL, returns to the previous phase; on PASS, advances and (at the validator gate) FREEZES the oracle. Writes phase transitions to the per-branch state file via the tdd-lib.sh CLI (the PreToolUse hook only reads). Refuses to start unless PRD+SPEC are active and the SPEC carries #### Scenario blocks (precondition C1). NEVER writes source/test/artifact files; NEVER calls forgeplan_activate. Cite ADR-010 contract C1-C6 + RFC-012 FR-7. RU: Мастер-оркестратор enforced-TDD под-цикла. Читает активные PRD+SPEC, затем диспатчит четыре фазы как отдельные изолированные контексты через Task — tdd-planner (сценарии→ПЛАН тестов) → coder-tdd (план→падающие тесты, RED) → tdd-test-validator (независимый C4-верификатор) → coder (GREEN; переиспользует agents-core:coder) — с БЛОКИРУЮЩИМ quality-gate между каждой фазой. На FAIL гейта возвращает к предыдущей фазе; на PASS продвигает и (на гейте валидатора) ЗАМОРАЖИВАЕТ оракул. Пишет переходы фаз в per-branch state-файл через CLI tdd-lib.sh (PreToolUse hook только читает). Отказывается стартовать, если PRD+SPEC не активны и SPEC не содержит блоков #### Scenario (precondition C1). НИКОГДА не пишет source/test/artifact файлы; НИКОГДА не вызывает forgeplan_activate. Цитирует ADR-010 contract C1-C6 + RFC-012 FR-7. Triggers: "tdd", "/tdd", "run tdd", "tdd cycle", "test-driven", "red green", "enforced tdd", "drive this spec with tests", "tdd orchestrator", "tdd master", "tdd-first feature", "запусти tdd", "tdd цикл", "красный зелёный", "разработка через тесты", "проведи через tdd"
Methodology: AD/AID-PDLC sub-cycle contract C3 (phase agent) + CRUD-R-A Profile A creator (test-PLAN artifact, language-NEUTRAL). First instance = enforced-TDD (RFC-012 FR-1 / ADR-010). EN: Language-NEUTRAL test planner — the first C3 phase agent of the enforced-TDD sub-cycle. Reads the FROZEN `#### Scenario` blocks of an active SPEC and produces a test PLAN: per scenario → test cases + what-to-assert + edge cases + RED-first expectation + how-to-write guidance for the downstream test-author. Writes NO code and picks NO language/framework. Emits the plan as a forgeplan artifact via MCP (never via Write/Edit). Denies source+test file writes and `forgeplan_activate`. May invoke `fpf:fpf-decompose` on a complex SPEC to split scenarios into bounded test groups (C7, on-demand). Hands the plan to `coder-tdd` (RED), who turns it into failing tests in the stack's engine. RU: Языково-НЕЙТРАЛЬНЫЙ планировщик тестов — первый C3 phase-агент enforced-TDD sub-cycle. Читает ЗАМОРОЖЕННЫЕ блоки `#### Scenario` активного SPEC и строит PLAN тестов: на каждый сценарий → тест-кейсы + что ассертить + edge cases + ожидание RED-first + как писать (для последующего автора тестов). НЕ пишет код и НЕ выбирает язык/фреймворк. Сохраняет план как forgeplan-артефакт через MCP (никогда через Write/Edit). Запрещает запись в исходники+тесты и `forgeplan_activate`. Может вызвать `fpf:fpf-decompose` на сложном SPEC, чтобы разбить сценарии на ограниченные группы тестов (C7, по требованию). Передаёт план `coder-tdd` (RED), который превращает его в падающие тесты в движке стека. Triggers: "plan the tests", "test plan from spec", "what to assert", "scenarios to test cases", "RED-first plan", "TDD plan", "tdd-planner", "what tests do we need", "распланируй тесты", "план тестов из спеки", "что ассертить", "сценарии в тест-кейсы", "что тестировать"
Methodology: AD/AID-PDLC C4 independent verifier (ADR-010) + CRUD-R-A Profile B-gate (binary PASS/CONCERNS/BLOCKER verdict). Implements RFC-012 FR-3. EN: The independent verifier for the TDD sub-cycle (Profile B-gate, guardian-shaped). The DIFFERENT context that certifies the RED tests before GREEN may begin — generator≠verifier applied to TESTS, not just code (ADR-009/ADR-010 C4). Reads the failing tests plus the SPEC's frozen `#### Scenario`s and certifies five things: every scenario has ≥1 covering test, the RED is valid (fails on an assertion, not a compile/collection error), the tests are not tautological/vacuous, assertion strength is adequate, and no mock gap hides a wiring failure. Renders a binary verdict — PASS (orchestrator may freeze the oracle and dispatch GREEN) / CONCERNS (orchestrator returns to coder-tdd to strengthen named tests) / BLOCKER (orchestrator halts; tests stay unfrozen) — and emits a forgeplan EVIDENCE artifact carrying that verdict and this agent's identity (the C6 exit gate reads it). It did NOT write the tests (coder-tdd did) — that separation is the whole point. Read-only: the `disallowedTools` denylist forbids ALL writes (Write/Edit/NotebookEdit) plus forgeplan_activate/reason. May invoke FPF `fpf-evaluate` (Trust Calculus) on contested coverage (C7). Never freezes the oracle itself and never activates — the orchestrator does both on reading a PASS verdict. RU: Независимый верификатор TDD sub-cycle (Profile B-gate, по форме guardian). ДРУГОЙ контекст, который сертифицирует RED-тесты до старта GREEN — generator≠verifier применён к ТЕСТАМ, а не только к коду (ADR-009/ADR-010 C4). Читает падающие тесты и замороженные `#### Scenario` из SPEC и сертифицирует пять вещей: каждый сценарий покрыт ≥1 тестом, RED валиден (падает на ассерте, а не на ошибке компиляции/сбора), тесты не тавтологичны/не вакуумны, сила ассертов адекватна, ни один мок-разрыв не скрывает ошибку проводки (wiring). Выносит бинарный вердикт — PASS (оркестратор может заморозить oracle и задиспатчить GREEN) / CONCERNS (оркестратор возвращает к coder-tdd усилить названные тесты) / BLOCKER (оркестратор останавливается; тесты остаются незамороженными) — и эмитит forgeplan EVIDENCE, несущий этот вердикт и identity этого агента (C6 exit gate читает его). Он НЕ писал тесты (это coder-tdd) — это разделение и есть весь смысл. Read-only: denylist `disallowedTools` запрещает ВСЕ записи (Write/Edit/NotebookEdit) плюс forgeplan_activate/reason. Может вызвать FPF `fpf-evaluate` (Trust Calculus) на спорном покрытии (C7). Сам никогда не замораживает oracle и не активирует — это делает оркестратор по прочтении PASS. Triggers: "validate the tests", "are the tests correct", "certify RED", "is this a valid red", "do the tests cover every scenario", "test quality gate", "vacuous green check", "проверь тесты", "тесты корректны", "сертифицируй RED", "валидный ли RED", "покрывают ли тесты все сценарии", "C4 verifier", "tdd test validator", "freeze the oracle", "tests cover scenarios", "tautological test check", "assertion strength check"
One-time per-project setup for the enforced-TDD methodology — generates the `.forgeplan/tdd/stack.json` language binding that the `tdd-gate.sh` hook reads (test command, test/source file globs, the assertion-failure marker, lint command). Without it the gate cannot classify files and the TDD cycle cannot start. Auto-detects the stack from repo markers (pyproject/go.mod/Cargo.toml/package.json/ composer.json/Gemfile/pom.xml), optionally reconciles with a stack-ADR if one exists, shows the result for human review, then writes it. EN: Run once before the first `/tdd` cycle on a project. Detects the test stack and writes `.forgeplan/tdd/stack.json`. Re-run to update after a stack change. RU: Запусти один раз перед первым циклом `/tdd` на проекте. Определяет тест-стек и пишет `.forgeplan/tdd/stack.json`. Перезапусти, чтобы обновить после смены стека. Triggers: "tdd init", "/tdd-init", "set up tdd", "tdd setup", "generate stack.json", "configure tdd", "заведи tdd", "настрой tdd", "stack.json для tdd"
Enforced-TDD entry point. Runs the C1-C6 sub-cycle the tdd-orchestrator drives: precondition check (PRD + SPEC active with #### Scenarios) → tdd-planner (scenarios → test plan) → coder-tdd (RED: plan → failing tests) → tdd-test-validator (C4: certify tests, freeze the SPEC oracle on PASS via a normalized full-file hash) → coder (GREEN: code to pass the frozen tests, cannot edit tests) → lint → EVIDENCE-out into the forgeplan Audit. Each tier is a separate isolated dispatch (generator≠verifier). Test immutability during GREEN is enforced by a fail-closed PreToolUse gate, not by prompt. Use when a feature has an active SPEC with #### Scenarios and you want tests-frozen-before-code with structural enforcement. Triggers: "tdd", "/tdd", "test-driven", "test driven development", "write tests first", "RED GREEN", "frozen oracle", "enforced tdd", "tests before code", "напиши тесты сначала", "TDD цикл", "красный зелёный"
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Official plugin marketplace for Claude Code from ForgePlan — UX, workflow, engineering, and developer tools.
18 plugins | 25 marketplace-aware agents (75 total) | 38+ skills | 9+ knowledge bases
ForgePlan ecosystem: this marketplace +
forgeplanCLI (artifact lifecycle) +@forgeplan/web(browser viewer with time-travel + graph). Three siblings; install what you need.
| Guide | When to read |
|---|---|
| 👋 Onboarding | Read this first — pocket guide, friendly tone, get from zero to autonomous agent dispatch in 30-60 min. Hindsight setup + command cookbook + when-you're-needed matrix |
| 🚧 Brownfield guide | Companion to Onboarding — existing codebases, Discover Agent, Strangler Fig + DDD + ACL with concrete examples |
| 🚀 Developer Journey | Longer 30-min walkthrough from zero to your first shipped feature, with 4 persona Day 0 walkthroughs |
| 🧭 Smith — master orchestrator | Routing map: 12 contexts × 27 methodologies. When and how to dispatch /smith as the BMAD-style master that picks the right workflow for the task |
| 📋 Playbook | Use-case matrix: which command for which scenario (empty project, brownfield, night-run, …) |
| 📖 Usage Guide | Reference manual: 18 commands, hooks, agent activation rules, troubleshooting |
| 🏛 Architecture | 4-layer mental model — Orchestra (where) · Forgeplan (what) · FPF (how to think) · SPARC (how to code) |
| 🔬 Methodologies | What's built into forgeplan (BMAD, OpenSpec, ADI, F-G-R, DDR, Verification Gate) vs what's external |
| 🔁 Autoresearch integration | Combining uditgoenka/autoresearch (metric-driven loops) with /forge-cycle lifecycle |
| 🧭 AI-SDLC mapping | Phase-by-phase map for users coming from AI-SDLC vocabulary (Concept → Research → Design → Build → Test → Release → Operate → Maintain) |
| 📚 Upstream methodologies | Bibliography — pointers to Quint-code, BMAD, OpenSpec, FPF, Karpathy autoresearch, git-adr, ccpm, adr-tools |
| 🛠 Setup Guide for new repo | End-to-end bootstrap (~20 min): forgeplan + Claude Code plugins + GitHub Projects v2 + auto-add workflow + convention. Use when adding a new ForgePlan repo to the ecosystem. |
| 🚚 Migration: dev-toolkit → fpl-skills | Switching plugins safely (side-by-side or clean-cut, with rollback plan) |
| 🗂 Tracker Integration | Recipes for Orchestra · GitHub Issues · Linear · Jira · local TODO |
| 🌐 Forgeplan Web | Browser viewer — time-travel slider + artifact graph + health dashboard |
| 🔄 Process Reference (EN) / RU | Full process from idea to commit — 5 agent profiles, 10 artifact kinds, 4-layer pipeline, methodologies (BMAD/FPF/SPARC/RIPER-5) |
| 📜 Changelog | Release history |
| 🤝 Contributing | Add a new plugin to the marketplace |
# 1. Add the ForgePlan marketplace
/plugin marketplace add ForgePlan/marketplace
# 2. Install your first plugin (works with any project, no dependencies)
/plugin install dev-toolkit@ForgePlan-marketplace
# 3. Reload to activate
/reload-plugins
# 4. Try it — run a code audit
/audit
New here? Pick the row that matches you. Each row is a complete setup — no need to read the full catalog.
| Your role | Install these | Why |
|---|---|---|
| I want a master orchestrator | fpl-skills + agents-pro | Start here → /smith routes 12 contexts to 27 methodologies |
| Forgeplan user (recommended) | fpl-skills | One install, 38 skills, full route → ship loop |
| Any developer (no forgeplan) | dev-toolkit + agents-core | Universal tools, no CLI dependency |
| Frontend | fpl-skills + laws-of-ux + agents-domain | UX + framework agents |
| Architect | fpl-skills + fpf + agents-pro + agents-sparc | Thinking + SPARC |
| Multi-session / team | fpl-skills + forgeplan-orchestra | Coordination + Inbox Pattern |
| Brownfield migration | fpl-skills + forgeplan-brownfield-pack | Ingest legacy docs + C4/DDD/research analyses |
| Everything | All 18 plugins | Complete ecosystem |
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 claimnpx claudepluginhub forgeplan/marketplace --plugin agents-tddForgePlan Hindsight memory plugin for Claude Code. Long-term cross-session memory wrapping Hindsight (vectorize-io). Provides 13 MCP tools (memory_* + mental_model_* + document_ingest_*), 3 auto-hooks (recall on UserPromptSubmit, retain on Stop, force-retain on SessionEnd), and 5 helper skills (/status, /bootstrap, /mental-model, /diagnose, /export-bank). Three activation modes: plugin install (default-on, bank derived from cwd), per-project setup CLI (explicit `.mcp.json`), or direct MCP wire-up (standalone bundle). Source code colocated under plugins/fpl-hsmem/src/. Implements Ruflo-style outcome-feedback patterns (NOTE-004).
Practical cookbook of forgeplan CLI recipes — task-oriented how-to guides covering getting-started, PRD authoring, Evidence patterns, workflow cycles, multi-plugin discipline, AI-pair dispatch, troubleshooting Sprint A-O anomalies, and CLI/MCP quick-reference. Each recipe: Цель → Команда → Пример → Common errors. / Практическая поваренная книга рецептов forgeplan CLI — пошаговые руководства по онбордингу, PRD, Evidence, рабочим циклам, мультиплагинной дисциплине, диспетчеризации AI-агентов, устранению аномалий и справочнику CLI/MCP.
UX laws knowledge base and frontend code reviewer. Analyzes HTML/CSS/JS/React/Vue/Svelte code against 30 Laws of UX principles. Provides actionable recommendations for improving user experience. Use when building UI components, reviewing frontend code, or learning UX best practices. v1.4.1: ux-reviewer agent canonical-lint compliant (LR-1..LR-3).
First Principles Framework (FPF) — thinking amplifier for structured reasoning, system decomposition, and decision-making. Enhanced version with applied patterns, forgeplan integration, and practical commands. Based on the FPF specification by Anatoly Levenchuk. v1.4.1: fpf-advisor agent canonical-lint compliant (LR-1..LR-3).
Unified workflow: Forgeplan artifacts + Orchestra task tracking + Claude Code AI execution. Bidirectional sync, Session Start Protocol with Inbox Pattern, and methodology knowledge base. Requires forgeplan CLI (private app, access via project admin) + Orchestra MCP server. v1.4.1: orchestra-advisor agent canonical-lint compliant (LR-1..LR-3).
Harness-native ECC plugin for engineering teams - 67 agents, 277 skills, 93 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
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.
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex and antigravity CLIs when installed) to get diverse perspectives on coding problems
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.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.