Discover MCPs & agents
Loading MCPs and agents…
Loading MCPs and agents…
AI Multi-Agent engineering team — one developer, full team power. Automates coding, testing, building, CI fixing, and code review across the entire dev lifecycle.
From the repo.
An AI Multi-Agent engineering team that empowers single developer = full engineering team
Automate the entire software development lifecycle: coding, testing, building, CI fixing, code review repair & comment resolution.
Traditional AI coding tools only generate fragmented code snippets. Dev Pilot runs a complete virtual engineering team:
It understands your project context, handles full-process engineering trivialities, and only leaves core architecture & decision work to humans.
The Dev Pilot Dashboard provides a visual command center for managing your AI engineering workflow.
Track open issues assigned to you, view status, and assign them to AI agents with one click.

Monitor active AI agent tasks in real-time — see progress bars, current phase, and elapsed time.

Manage open PRs authored by your AI team — review status, merge targets, and trigger reviews.

Get an end-of-day summary: issues closed, PRs merged, and a timeline of completed work.

git clone https://github.com/frankliu20/dev-pilot.git && cd dev-pilot
node init.js # Prompts for workspace path + GitHub repo (owner/repo)
init.js creates ~/.claude/pilot.yaml with your settings and installs commands/agents. Review and tweak it anytime — see the Configuration Guide.
# Optional: start the dashboard
cd dashboard && npm install && npm run dev
Then, in Claude Code:
/pilot-dev-issue https://github.com/your-org/your-repo/issues/123
That's it. Sit back and watch, or go work on something else.
# Fire and forget — multiple issues in parallel, fully unattended
claude "/pilot-dev-issue --auto https://github.com/org/repo/issues/123" &
claude "/pilot-dev-issue --auto https://github.com/org/repo/issues/456" &
claude "/pilot-dev-issue --auto https://github.com/org/repo/issues/789" &
| Component | What it does |
|---|---|
/pilot-dev-issue | The core pipeline: analyze → explore → plan → code → test → PR |
/pilot-watch-pr | Monitors open PRs, auto-fixes CI failures, notifies on reviews |
pilot-code-explorer | Parallel codebase analysis agents (2-3 per issue) |
pilot-pr-creator | Handles git operations: stage, commit, push, create PR |
pilot-pr-reviewer | Structured code review with interactive discussion |
| Dashboard | Web UI at localhost:3000 — issues, PRs, tasks, decisions at a glance |
You
│
├── /pilot-dev-issue #123
│ ├── Analyze issue
│ ├── Explore codebase (parallel agents)
│ ├── Present plan → you approve
│ ├── Implement + test + auto-fix
│ └── Open PR
│
├── /pilot-watch-pr
│ ├── Poll PRs every 5 min
│ ├── Auto-fix CI failures
│ └── Notify when ready to merge
│
└── Dashboard (localhost:3000)
└── Visual command center for everything above
gh) — run gh auth login after install| Document | Description |
|---|---|
| Configuration Guide | Setup pilot.yaml, permissions, build/test commands |
| Architecture | Project structure and data flow |
| Developer Guide | Build, test, and contribute |
| Contributing | How to contribute to Dev Pilot |
MIT — see LICENSE.
Replace {MCP_ENDPOINT_URL} with this MCP’s endpoint URL (from its repo or docs above). No API key — you connect directly.
Tool
OS
Config file: ~/.cursor/mcp.json
{
"mcpServers": {
"mcp-server": {
"url": "{MCP_ENDPOINT_URL}"
}
}
}Paste into mcpServers in the config file. Restart Cursor after saving.
If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.