Manages PLAN.md through MCP tools for agent-readable progress, scoped task updates, iteration lifecycle, and plan validation.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent β or use 1-click editor setup below.
We haven't yet run this listing's install command through our automated sandbox check. This isn't a red flag β we're steadily working through the catalog.
π‘ Paste the JSON block into your client's configuration file under mcpServers, then restart the application.
Inspect callable tools, capabilities, and parameters exposed to AI agents by Powerplan.
create_planBootstrap `./PLAN.md` (or `plan_path`) when missing; `force` to overwrite
show_miniplanSession opener** β raw PLAN.md snippet: the current (or named) iteration byte-for-byte, neighbours collapsed to header lines (`before`/`after`)
get_current_iterationPreferred for agents** β scoped JSON for current work
get_iterationJSON for one version (tasks, progress)
list_iterationsNavigate without full-file reads
create_majorSurgical mutations (batch add in one write)
The powerplan MCP server turns a projectβs PLAN.md into an MCP-accessible planning interface. It is intended for coordinator and worker agents that need to inspect current work, update tasks, and keep iteration status consistent while avoiding unrestricted file edits.
The managed format recognizes major versions such as ## vX.Y β Title, iterations such as ### vX.Y.Z β Title, a goal line, checkbox tasks, and a backlog section. Other phase-like headings and prose are preserved as opaque content. The server can also add optional trailing [agent: id] markers to lines changed by mutations.
Use it when PLAN.md is the shared operational record for a project or agent group. It can be paired with PowerSpawn, although the README specifies that the two MCP servers remain separately registered and do not merge.
The server communicates over MCP using stdio. By default, tools walk upward from the current working directory to find the nearest PLAN.md. Each tool also accepts an optional relative or absolute plan_path, allowing callers to select a different file.
A typical agent session starts with show_miniplan, which returns the active or selected iteration byte-for-byte while reducing neighboring iterations to header lines. get_current_iteration provides the preferred scoped JSON representation for agents, and get_iteration returns tasks and progress for one version. Navigation tools such as list_iterations, find_task, and get_backlog avoid requiring a full-file read.
Mutations are performed through dedicated operations rather than freeform file replacement. The server writes changes surgically and can batch task additions or completions. check_plan checks plan structure, while show_plan and show_current_iteration provide compact human-oriented views.
Install with uv and run the stdio server using:
The package requires uv, which provides uvx, or Python 3.10 or newer. Without uv, install the PyPI package with pip install powerplan-mcp, then launch it with python -m powerplan.
Clients can register the process with the powerplan-mcp command and pass PYTHONIOENCODING=utf-8 and PYTHONUNBUFFERED=1 in the environment, as shown in the README examples. Explicit configurations are provided for Claude Code, Cursor, Claude Desktop, and Grok. A source checkout can be installed in editable mode with pip install -e ".[dev]", run with python -m powerplan, or started directly through powerplan_server.py.
The powerplan MCP server includes tools for:
complete_task accepts indexes or task selections, and task mutations can include an agent identifier. If a tool reports that PLAN.md is missing, create the file first with create_plan.
The server is focused on PLAN.md rather than general project-file editing. show_miniplan is designed as a session opener, and show_plan is a compact human skim; neither is described as a complete file dump. The README also advises agents not to read all of PLAN.md just to determine their next action.
The default path behavior depends on the process working directory, so callers operating outside the project tree should pass plan_path. A force-enabled create_plan can overwrite an existing plan, so it should be used deliberately. The published package name is powerplan-mcp; powerplan on PyPI refers to an unrelated package.
Factual signals from GitHub, npm, and our automated checks β not a rating.
No reviews yet β be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/powerplan)<a href="https://allmcps.com/mcp/powerplan"><img src="https://allmcps.com/api/badge/powerplan?style=directory" alt="Powerplan on AllMCPs" /></a>