By tctinh
Orchestrate plan-first AI development with parallel autonomous agents in isolated git worktrees. Plan multi-step tasks, enforce TDD for implementations, dispatch foragers for bugs and features, execute in batched sessions with checkpoints, verify builds/tests/lints, and run systematic code reviews before merging.
Autonomous worker executing tasks in isolated worktrees. Writes code, runs tests, reports results via hive_worktree_commit.
Plan-first AI development orchestrator. Plans features, dispatches parallel workers in isolated worktrees, merges results, coordinates reviews. The user's primary interface for Hive workflow.
Falsification-first code reviewer. Challenges implementation against plan and specs. Runs tests before giving verdicts. Opus tier cross-checks Sonnet workers.
Use when bootstrapping, updating, or reviewing AGENTS.md — teaches what makes effective agent memory, how to structure sections, signal vs noise filtering, and when to prune stale entries
Use before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
Use when reviewing implementation changes against an approved plan or task (especially before merging or between Hive tasks) to catch missing requirements, YAGNI, dead code, and risky patterns
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
Use when working with Docker containers — debugging container failures, writing Dockerfiles, docker-compose for integration tests, image optimization, or deploying containerized applications
Uses power tools
Uses Bash, Write, or Edit tools
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.
Plan first. Execute with trust. Context persists.
Agent Hive is a workflow layer that sits on top of your AI coding tool. It imposes just enough structure to make multi-agent, multi-step work traceable and recoverable — without taking ownership of your editor, your model, or your coding style.
https://github.com/user-attachments/assets/6290b435-1566-46b4-ac98-0420ed321204
Raw agentic coding has a consistent failure mode: agents spray changes across a codebase, sessions lose context, parallel workers collide, and nobody can reconstruct what happened. Hive fixes this with a small, strict loop:
You describe the work
↓
Hive discovers, asks, builds plan.md
↓
You review and approve ← human gate
↓
Workers execute tasks in isolated git worktrees (batched parallel)
↓
Results merge. plan/spec/report live in .hive/ forever.
| Without Hive | With Hive |
|---|---|
| Agent touches 40 files, half break | Tasks run in isolated worktrees — discard any worker |
| New session starts from zero | Feature state persists in .hive/features/<name>/ |
| Parallel agents collide, duplicate | Explicit batches with dependency ordering |
| "What happened here?" | plan.md, spec.md, report.md per task |
| Scope creep mid-execution | Human approval gate before any code change |
| Platform | Hive role | Best for |
|---|---|---|
| Claude Code | Full runtime (plugin + MCP) | Claude Code CLI users |
| OpenCode | Full runtime (plugin) | OpenCode CLI users — most feature-complete |
| VS Code | Companion UI (review + status) | Visual review alongside a CLI runtime |
The VS Code extension is not a runtime. It visualises the .hive/ state written by Claude Code or OpenCode and lets you approve plans without leaving the editor.
The Claude Code plugin ships three Hive agents (hive, forager, hygienic), a /hive slash command, 11 on-demand skills, a SessionStart hook, and spawns the @tctinh/agent-hive-mcp MCP server as a sidecar.
# inside Claude Code
/plugin marketplace add tctinh/agent-hive
/plugin install hive@agent-hive
The marketplace definition lives at .claude-plugin/marketplace.json and points at packages/claude-code-hive. The MCP runtime is loaded via node ./scripts/launch-hive-mcp.mjs, which resolves @tctinh/agent-hive-mcp from a node_modules tree. For marketplace installs, install the MCP runtime once so the launcher can resolve it:
npm install -g @tctinh/agent-hive-mcp
Useful when you prefer a single project-local dependency tree and don't want a global install:
mkdir -p .agent-hive/claude && cd .agent-hive/claude
npm init -y
npm install claude-code-hive @tctinh/agent-hive-mcp
Then register the plugin by pointing Claude Code at .agent-hive/claude/node_modules/claude-code-hive/.claude-plugin/plugin.json.
/hive add user authentication
/hive is the only entry point you need — it creates a feature, runs discovery, writes the plan, waits for approval, and dispatches workers.
npx claudepluginhub tctinh/agent-hive --plugin hiveMulti-agent orchestration for code that matters.
Long-running agent harness with 5-layer memory architecture, GitHub integration, autonomous batch processing, Agent Teams with ATDD, 9 hooks (safety, quality gates, team coordination), and 6 Agent Skills
Long Task Harness for AI agents - task/feature-driven development with external memory
Implementation planning, execution, and PR creation workflows with multi-agent collaboration
Persona-driven AI development team: orchestrator, team agents, review agents, skills, slash commands, and advisory hooks for Claude Code
OpenAgentsControl — multi-agent orchestration for Claude Code. Context-aware development with skills, subagents, parallel execution, and automated code review.