Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

AgentHub — Multi-Agent Competition Hub

Spawns N parallel subagents that solve the same task in isolated git worktrees, then evaluates and merges the winning branch.

AutomationAdvanced★ 24,151⑂ 3,405AI score 8/10Last updated: Aug 9, 2026

What it does

AgentHub turns a single task into a tournament between N parallel AI agents.

  • /hub:init creates a session (task, agent count, eval criteria); /hub:spawn launches all agents in one parallel batch.
  • Each agent works independently in its own git worktree on hub/{session}/agent-{N}, with no visibility into other agents' work.
  • /hub:eval ranks results via metric mode (benchmarks, test pass rates), LLM judge (correctness, simplicity, quality), or a hybrid tie-breaker.
  • /hub:merge merges the winner with --no-ff and tags the losers for archival — the DAG is append-only, so every attempt is preserved.
  • Ships iteration templates (optimizer, refactorer, test-writer, bug-fixer), a message board under .agenthub/board/, and helper scripts for DAG analysis, ranking, and session cleanup.

Who it's for

  • Engineers who want several optimization or bug-fix approaches tried at once and picked by hard numbers.
  • Teams comparing refactoring strategies before committing to one.
  • Content folks generating multiple drafts or A/B copy variations.
  • Intermediate-to-advanced users comfortable with git worktrees (a git repo is required).

Examples

  1. Performance bake-off: "Have 5 agents optimize this sort path and merge the lowest p50" — each worktree runs the benchmark, and 142ms beats the 180ms baseline.
  2. Refactor comparison: spawn 3 agents with --template refactorer, keep only branches where tests pass, and let the LLM judge pick the smallest clean diff.
  3. Copy A/B test: generate 4 landing-page copy variants in parallel, rank them on clarity and CTA strength, and merge only the winner.

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the .gemini/skills/agenthub folder from the GitHub repo alirezarezvani/claude-skills into my ~/.claude/skills/agenthub/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/alirezarezvani/claude-skills.git && mkdir -p ~/.claude/skills && cp -r claude-skills/.gemini/skills/agenthub ~/.claude/skills/agenthub

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal in a scratch directory.
  2. Clone the repo: git clone https://github.com/alirezarezvani/claude-skills.git
  3. Create the skills folder: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r claude-skills/.gemini/skills/agenthub ~/.claude/skills/agenthub
  5. Verify SKILL.md and the scripts/ folder exist; make sure Python 3 is installed since the helper scripts need it.
  6. Restart Claude Code, open a project that is a git repository, and run /hub:init.
  7. For your first run, use 2–3 agents and try /hub:run to walk the full lifecycle (init → spawn → eval → merge).
  8. Always review /hub:status and git diff before letting it merge a winning branch.
View source on GitHub ↗License: MIT