{
  "markdown": "# 5dive MCP server\n\n[![npm](https://img.shields.io/npm/v/@5dive/mcp)](https://www.npmjs.com/package/@5dive/mcp)\n[![Awesome MCP Servers](https://img.shields.io/badge/Awesome-MCP%20Servers-8A2BE2)](https://github.com/punkpeye/awesome-mcp-servers)\n\nExpose the [**5dive**](https://5dive.ai) agent-fleet CLI as [Model Context\nProtocol](https://modelcontextprotocol.io) tools. Point any MCP client (Claude\nDesktop, Cursor, Cline, or your own) at this stdio server to file tasks, inspect\nand message agents, and read the fleet digest — directly from inside a model\ncontext.\n\n[5dive](https://5dive.ai) is the CLI + control plane for running a fleet of\nautonomous coding agents as a self-governing company. This server is a thin,\nhonest adapter: every tool shells out to the local `5dive` binary's\nmachine-readable `--json` surface and returns the result — so it inherits the\nCLI's auth, permissions, and audit log for free, and never handles secrets itself.\n\n## Tools\n\n| Tool | Wraps | What it does |\n| --- | --- | --- |\n| `task_create` | `5dive task add` | File a task in the shared queue (title, body, priority, assignee, parent). |\n| `task_show` | `5dive task show` | Full detail for one task by id (status, body, result, subtasks, blockers). |\n| `task_list` | `5dive task ls` | List tasks (open by default; filter by status / assignee). |\n| `agent_send` | `5dive agent send` | Send a message to another agent on the fleet. |\n| `agent_list` | `5dive agent list` | List every agent: type, channels, model, live state. |\n| `digest_get` | `5dive digest` | Fleet daily standup digest (`window: \"7d\"` for the weekly view). |\n\n## Requirements\n\n- Node.js >= 18\n- The `5dive` CLI installed and on `PATH` (`curl https://install.5dive.ai | sudo bash`).\n\n## Install & run\n\n```bash\nnpx @5dive/mcp        # run directly\n# or\nnpm i -g @5dive/mcp && 5dive-mcp\n```\n\n### Client config (Claude Desktop / Cursor / Cline)\n\n```json\n{\n  \"mcpServers\": {\n    \"5dive\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@5dive/mcp\"],\n      \"env\": { \"FIVEDIVE_SUDO\": \"1\" }\n    }\n  }\n}\n```\n\n## Configuration (env vars)\n\n| Var | Default | Purpose |\n| --- | --- | --- |\n| `FIVEDIVE_BIN` | `5dive` | Path to the 5dive binary. |\n| `FIVEDIVE_SUDO` | *(unset)* | Set to `1` to prefix calls with `sudo`. Managed 5dive boxes require root for most subcommands; leave unset if you already run as root. |\n| `FIVEDIVE_TIMEOUT_MS` | `30000` | Per-call timeout in milliseconds. |\n\n## Safety\n\nArguments are passed to the CLI as an argv array with **no shell**, so tool input\ncan never be interpreted as shell syntax. The server never sees secrets: the CLI\nreads its own credentials from the box.\n\n## Scope\n\nThis mirrors a curated slice of the CLI (tasks, agents, digest), not its full\nsurface. It is a distribution and convenience layer, not a new API. For\neverything else, use the `5dive` CLI directly (`5dive --help`). Full docs: https://5dive.ai/docs/5dive-cli\n\n## License\n\nMIT\n",
  "bytes": 2931,
  "sha": "755cc95e2ed615bc6169b3b4aee15ab6d4dc24ec1e1749f9751984ce063da954",
  "repo_slug": "5dive-ai/5dive-mcp",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_5dive_ai_5dive_mcp_3488e3b3/readme"
}