deepeshBodh/human-in-loop

Última actividad de origen registrada: 17 de marzo de 2026 a las 06:01

Catálogo de SkillsMP actualizado: 12 de noviembre de 2026 a las 12:52

skills recopiladas: 30

Estrellas en GitHub: 39
Forks en GitHub: 2

Instalar con un asistente de IA

Copia este prompt en el asistente de IA que estés usando.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/deepeshBodh/human-in-loop.

README del repositorio

Archivo de origen del README: README.md

Leer el README original

Mostrar README completo

# HumanInLoop **Stop vibe coding. Ship software that lasts.** [Website](https://humaninloop.dev) • [Quick Start](#quick-start) • [Roadmap](./ROADMAP.md) • [Changelog](./CHANGELOG.md) --- ## What is HumanInLoop? HumanInLoop is a Claude Code plugin that enforces **specification-driven development**—ensuring architectural decisions are made by humans before AI writes code. Instead of letting AI improvise your architecture, you guide it through a structured workflow: ``` Idea → Specification → Plan → Tasks → Implementation ``` Every step produces artifacts you can review, refine, and approve before moving forward. --- ## The Workflow ``` ┌──────────────────────────────────────────────────────────────────────────────────┐ │ │ │ ┌─────────┐ ┌─────────┐ ┌──────┐ ┌───────┐ ┌───────────┐ │ │ │ SETUP │─▶│ SPECIFY │─▶│ PLAN │─▶│ TASKS │─▶│ IMPLEMENT │ │ │ └─────────┘ └─────────┘ └──────┘ └───────┘ └───────────┘ │ │ │ │ │ │ │ │ │ ▼ ▼ ▼ ▼ ▼ │ │ ┌─────────┐ ┌─────────┐ ┌──────┐ ┌───────┐ ┌───────────┐ │ │ │ consti- │ │ spec │ │ 6 │ │ tasks │ │ code │ │ │ │ tution │ │ .md │ │ files│ │ .md │ │ changes │ │ │ └─────────┘ └─────────┘ └──────┘ └───────┘ └───────────┘ │ │ │ │ │ │ │ │ │ └────────────┴───────────┴──────────┴────────────┘ │ │ │ │ │ ┌─────▼─────┐ │ │ │ AUDIT │ │ │ │ (review) │ │ │ └───────────┘ │ │ │ │ ○────────○────────○────────○────────○ Human review checkpoints │ │ │ └──────────────────────────────────────────────────────────────────────────────────┘ ``` | Stage | Command | What You Get | |-------|---------|--------------| | **Setup** | `/humaninloop:setup` | Project constitution with your standards | | **Specify** | `/humaninloop:specify` | Structured spec with user stories and requirements | | **Plan** | `/humaninloop:plan` | Technical requirements, constraints, decisions, data models, API contracts | | **Tasks** | `/humaninloop:tasks` | Ordered task list with TDD cycles | | **Audit** | `/humaninloop:audit` | Quality analysis across all artifacts | | **Implement** | `/humaninloop:implement` | Guided implementation with progress tracking | Each command produces artifacts you review before the next step. You stay in control. ### Command Details <details open> <summary>Specify - Create feature specification (DAG-based)</summary> ``` ┌─────────────────────────────────────────────────────────────────────┐ │ /humaninloop:specify │ ├─────────────────────────────────────────────────────────────────────┤ │ │ │ ┌──────────────┐ │ │ │ SUPERVISOR │ ◄─── You invoke the command │ │ └──────┬───────┘ │ │ │ ┌──────────────────┐ ┌──────────────────┐ │ │ ├─▶│ State Analyst │────▶│ briefing + │ │ │ │ │ (analysis + │ │ assembly + │ │ │ │ │ graph ops) │ │ recommendations │ │ │ │ └──────┬───────────┘ └────────┬─────────┘ │ │ │ │ │ │ │ │ ▼ ▼ │ │ │ ┌──────────────────┐ ┌───────────────────┐ │ │ │ │ Requirements │────▶│ Devil's │ │ │ │ │ Analyst │ │ Advocate │ │ │ │ └────────┬─────────┘ └─────────┬─────────┘ │ │ │ │ │ │ │ │ ▼ ▼ │ │ │ ┌─────────┐ ┌────────────────────┐ │ │ │ │ spec.md │ │ gaps? → new pass │ │ │ │ └─────────┘ └────────────────────┘ │ │ │ │ │ ▼ Deterministic DAG tracks all nodes, edges, and status │ │ ┌──────────────────────────────────────────────────────────┐ │ │ │ StrategyGraph JSON (single DAG, multi-pass iteration) │ │ │ └──────────────────────────────────────────────────────────┘ │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` **Output**: `specs/{feature}/spec.md` + StrategyGraph JSON **Agents**: - **State Analyst** — Strategic analysis + graph mechanics; produces briefings, assembles nodes, freezes passes, and constructs domain agent prompts - **Requirements Analyst** — Transforms feature requests into structured specs; no implementation details - **Devil's Advocate** — Reviews for gaps and ambiguity; asks clarifying questions </details> <details> <summary>Plan - Unified analysis and design planning</summary> ``` ┌─────────────────────────────────────────────────────────────────────┐ │ /humaninloop:plan │ ├─────────────────────────────────────────────────────────────────────┤ │ │ │ ┌──────────────┐ │ │ │ SUPERVISOR │ ◄─── Reads spec.md │ │ └──────┬───────┘ │ │ │ │ │ ▼ Phase P1 (Analysis) │ │ ┌──────────────────┐ ┌───────────────────┐ ┌─────────────────┐ │ │ │ Technical │─▶│ Principal │─▶│ Devil's │ │ │ │ Analyst │ │ Architect │ │ Advocate │ │ │ └────────┬─────────┘ └───────────────────┘ └────────┬────────┘ │ │ │ (feasibility gate) │ │ │ ▼ ▼ │ │ ┌──────────────┐ ┌───────────────────┐ ┌─────────┐ ┌────────┐ │ │ │requirements │ │constraints-and- │ │ nfrs.md │ │ gaps? │ │ │ │ .md │ │ decisions.md │ │ │ └────────┘ │ │ └──────────────┘ └───────────────────┘ └─────────┘ │ │ │ │ │ Phase P2 (Design) │ │ ▼ │ │ ┌──────────────────┐ ┌─────────────────┐ │ │ │ Technical │────▶│ Devil's │ │ │ │ Analyst │ │ Advocate │ │ │ └────────┬─────────┘ └────────┬────────┘ │ │ │ │ │ │ ▼ ▼ │ │ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │ │ │ data- │ │contracts/│ │quickstart│ │ gaps? │ │ │ │ model.md │ │ api.yaml │ │ .md │ └──────────┘ │ │ └──────────┘ └──────────┘ └──────────┘ │ │ │ │ │ │ │ └──────────┴───────────┘ │ │ │ │ │ ▼ │ │ ┌──────────┐ │ │ │ plan.md │ │ │ └──────────┘ │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` **Output**: `specs/{feature}/plan.md` + 6 supporting artifacts **Agents**: - **Technical Analyst** — Produces analysis artifacts (requirements, constraints-and-decisions, NFRs) and design artifacts (data model, API contracts, integration guide) - **Principal Architect** — Reviews cross-artifact feasibility after analysis phase (one-time gate) - **Devil's Advocate** — Validates completeness, traceability, and cross-artifact consistency </details> <details> <summary>Tasks - Generate implementation tasks</summary> ``` ┌─────────────────────────────────────────────────────────────────────┐ │ /humaninloop:tasks │ ├─────────────────────────────────────────────────────────────────────┤ │ │ │ ┌──────────────┐ │ │ │ SUPERVISOR │ ◄─── Reads plan artifacts │ │ └──────┬───────┘ │ │ │ │ │ ▼ │ │ ┌──────────────────┐ ┌───────────────────┐ │ │ │ Task │────▶│ Devil's │ │ │ │ Architect │ │ Advocate │ │ │ └────────┬─────────┘ └─────────┬─────────┘ │ │ │ │ │ │ ▼ ▼ │ │ ┌────────────────┐ ┌──────────┐ │ │ │ task-mapping.md│ │ gaps? │──── yes ──┐ │ │ └────────────────┘ └──────────┘ │ │ │ │ │ ▼ │ │ ▼ no ┌────────────┐ │ │ ┌──────────┐ │ │ clarify w/ │ │ │ │ tasks.md │ ▼ │ user │ │ │ └──────────┘ ┌──────────┐ └─────┬──────┘ │ │ │ done │ │ │ │ └──────────┘ ◄─────┘ (loop) │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` **Output**: `specs/{feature}/tasks.md` with TDD cycles **Agents**: - **Task Architect** — Maps requirements to vertical slices with TDD structure; no implementation - **Devil's Advocate** — Ensures coverage, proper ordering, and testable increments </details> <details> <summary>Setup - Create project constitution</summary> ``` ┌─────────────────────────────────────────────────────────────────────┐ │ /humaninloop:setup │ ├─────────────────────────────────────────────────────────────────────┤ │ │ │ ┌──────────────┐ │ │ │ SUPERVISOR │ │ │ └──────┬───────┘ │ │ │ │ │ ▼ │ │ ┌──────────────────┐ Brownfield? │ │ │ Detect Codebase │────────┬──────────────┐ │ │ └──────────────────┘ │ │ │ │ yes no │ │ │ │ │ │ ▼ │ │ │ ┌──────────────────┐ │ │ │ │ Principal │ │ │ │ │ Architect │ │ │ │ └────────┬─────────┘ │ │ │ │ │ │ │ ┌───────────────────┼───────────────┤ │ │ ▼ ▼ ▼ │ │ ┌────────────┐ ┌────────────────┐ ┌────────────┐ │ │ │ codebase- │ │ constitution │ │constitution│ │ │ │analysis.md │ │ .md │ │ .md │ │ │ └────────────┘ └────────────────┘ └────────────┘ │ │ │ │ │ │ ▼ │ │ │ ┌────────────┐ │ │ │ │ evolution- │ │ │ │ │ roadmap.md │ │ │ │ └────────────┘ │ │ │ └───────────────────┘ │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` **Output**: `.humaninloop/memory/constitution.md` **Agents**: - **Principal Architect** — Defines governance principles and quality gates; enforces RFC 2119 keywords </details> <details> <summary>Audit - Analyze artifacts for quality</summary> ``` ┌─────────────────────────────────────────────────────────────────────┐ │ /humaninloop:audit │ ├─────────────────────────────────────────────────────────────────────┤ │ │ │ ┌──────────────┐ │ │ │ SUPERVISOR │ ◄─── No agents, direct analysis │ │ └──────┬───────┘ │ │ │ │ │ ▼ │ │ ┌──────────────────────────────────────────────────────────┐ │ │ │ Load Artifacts │ │ │ │ spec.md │ plan.md │ tasks.md │ data-model │ contracts │ │ │ └──────────────────────────┬───────────────────────────────┘ │ │ │ │ │ ┌───────────────────┼───────────────────┐ │ │ ▼ ▼ ▼ │ │ ┌────────────┐ ┌────────────┐ ┌────────────┐ │ │ │ Spec │ │ Plan │ │ Task │ │ │ │ Analysis │ │ Validation│ │ Validation │ │ │ └────────────┘ └────────────┘ └────────────┘ │ │ │ │ │ │ │ └───────────────────┴───────────────────┘ │ │ │ │ │ ▼ │ │ ┌────────────────┐ │ │ │ Audit Report │ │ │ │ (stdout/file) │ │ │ └────────────────┘ │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` **Output**: Coverage report + flagged issues **Agents**: None — direct analysis using validation skills </details> <details> <summary>Implement - Execute the task plan</summary> ``` ┌─────────────────────────────────────────────────────────────────────┐ │ /humaninloop:implement │ ├─────────────────────────────────────────────────────────────────────┤ │ │ │ ┌──────────────┐ │ │ │ SUPERVISOR │ ◄─── Reads tasks.md │ │ └──────┬───────┘ │ │ │ │ │ ▼ │ │ ┌──────────────────────────────────────────────────────────┐ │ │ │ Foundation Cycles (Sequential) │ │ │ │ C1 ──▶ C2 ──▶ C3 ──▶ ... │ │ │ └──────────────────────────┬───────────────────────────────┘ │ │ │ │ │ ▼ │ │ ┌──────────────────────────────────────────────────────────┐ │ │ │ Feature Cycles (Can Parallel) │ │ │ │ C4 ──┬──▶ C5 │ │ │ │ └──▶ C6 [P] │ │ │ └──────────────────────────┬───────────────────────────────┘ │ │ │ │ │ For each cycle: │ │ │ ┌──────────────────────────┴───────────────────────────────┐ │ │ │ T1 (test) ──▶ T2 (impl) ──▶ T3 (impl) ──▶ Checkpoint │ │ │ │ │ │ │ │ │ ▼ (if TEST: marker) │ │ │ │ ┌─────────────────┐ │ │ │ │ │ QA Engineer │ ──▶ auto-approve or human checkpoint│ │ │ │ └─────────────────┘ │ │ │ └──────────────────────────────────────────────────────────┘ │ │ │ └─────────────────────────────────────────────────────────────────────┘ ``` **Output**: Implemented code + marked tasks **Agents**: - **QA Engineer** — Executes verification tasks and captures evidence; auto-approves or presents checkpoints </details> --- ## What's Included ### 6 Commands The full specify → plan → tasks → implement lifecycle. ### 29 Skills Claude automatically invokes these when relevant—authoring requirements, technical specifications, analyzing codebases, designing APIs, running verification tests, managing GitHub issues, workflow strategy, and more. ### 9 Specialized Agents Focused responsibilities: requirements analyst, technical analyst, devil's advocate, principal architect, task architect, staff engineer, QA engineer, UI designer, state analyst. See the [plugin documentation](./plugins/humaninloop/README.md) for full details. --- ## Quick Start ### 1. Add the marketplace ```bash /plugin marketplace add deepeshBodh/human-in-loop ``` ### 2. Install the plugin ```bash /plugin install humaninloop ``` ### 3. Install and configure the `hil-dag` MCP server (required for specify and implement) ```bash uv tool install "humaninloop-brain @ git+https://github.com/deepeshBodh/human-in-loop.git#subdirectory=humaninloop_brain" ``` Then configure the MCP server in your Claude Code settings (`.claude/settings.json`): ```json { "mcpServers": { "hil-dag": { "command": "hil-dag" } } } ``` **To upgrade to the latest version:** ```bash uv tool install --force "humaninloop-brain @ git+https://github.com/deepeshBodh/human-in-loop.git#subdirectory=humaninloop_brain" ``` ### 4. Set up your project ```bash /humaninloop:setup ``` This creates your project constitution—the standards and conventions that guide all future specifications. ### 5. Create your first spec ```bash /humaninloop:specify add user authentication with email and password ``` --- ## Documentation | Resource | Description | |----------|-------------| | [Roadmap](./ROADMAP.md) | Vision and planned features | | [Changelog](./CHANGELOG.md) | Release history | | [Plugin README](./plugins/humaninloop/README.md) | Detailed command and skill reference | --- ## For Plugin Developers This repository serves as a reference implementation for Claude Code plugins. If you're building your own plugins, you can learn from: ### Repository Structure ``` human-in-loop/ ├── humaninloop_brain/ # Deterministic DAG infrastructure (Python) │ ├── src/humaninloop_brain/ # Package source │ │ ├── entities/ # Pydantic models (11 enums, 14 models) │ │ ├── graph/ # NetworkX graph operations │ │ ├── validators/ # Structural + contract validators │ │ ├── passes/ # Pass lifecycle management │ │ ├── mcp/ # MCP server + transport-agnostic operations │ │ └── cli/ # CLI adapter (delegates to mcp/operations) │ └── tests/ # 403 tests, ~95% coverage ├── plugins/humaninloop/ │ ├── .claude-plugin/plugin.json # Plugin manifest │ ├── commands/ # Slash command definitions │ ├── agents/ # 9 specialized agent definitions │ ├── skills/ # 29 model-invoked skills │ ├── catalogs/ # Node catalogs for DAG workflows │ ├── templates/ # Workflow templates │ └── scripts/ # Shell utilities ├── docs/ │ ├── decisions/ # Architecture Decision Records (8 ADRs) │ ├── architecture/ # DAG-first + v3 architecture docs │ ├── claude-plugin-documentation.md │ └── agent-skills-documentation.md └── specs/ # Feature specifications (dogfooding) ``` ### Key Resources - [Claude Code Plugin Documentation](./docs/claude-plugin-documentation.md) - Complete technical reference - [Agent Skills Documentation](./docs/agent-skills-documentation.md) - How skills work - [Architecture Decisions](./docs/decisions/) - ADRs explaining design choices --- ## Contributing See [CONTRIBUTING.md](./CONTRIBUTING.md) for guidelines. --- ## License MIT - See [LICENSE](./LICENSE)

Cargando el README del repositorio…

Skills en este repositorio

Desarrolladores de software Analistas de garantía de calidad de software y probadores Analistas de sistemas informáticos Diseñadores gráficos

13 categorías ocupacionales · 100% clasificado

skill

ocupación

descripción

actualizado

strategy-implementation

Especialistas en gestión de proyectos

This skill MUST be invoked when the user says "implementation strategy", "cycle sequencing", or "implement workflow patterns". SHOULD also invoke when user mentions "execute-then-verify", "targeted retry", "fix pass", or "implementation escalation". Provides…