hacknplan-mcp

hacknplan-mcp

Enables Claude to interact with HacknPlan project management via the MCP protocol. Provides tools for managing projects, work items, boards, milestones, time logs, design models, and cross-project portfolio views.

Category
Visit Server

README

@mtuska/hacknplan-mcp

A Model Context Protocol (MCP) server for the HacknPlan project-management API, built for use with Claude. It exposes HacknPlan projects, boards, work items, sub-tasks, dependencies, time logs, the design model, and metrics as MCP tools — plus a cross-project portfolio dashboard and a deadline schedule view that HacknPlan has no native equivalent for.

Pure TypeScript/Node — runs over npx, no Python or other runtime required.

Quick start

npx @mtuska/hacknplan-mcp install

The installer will:

  1. Ask where to register the server with Claude — global (all your projects), project (this repo, shared via .mcp.json), or local (this repo, private to you).
  2. Prompt for your HacknPlan API key (input is hidden) and store it in the chosen Claude MCP config.
  3. Install the bundled hacknplan Claude skill into the matching skills directory (global → ~/.claude/skills, project/local → ./.claude/skills).

Then ask Claude to run the hacknplan_whoami tool to confirm the key works. (Restart Claude / reload the window first if it was already running.)

The bundled skill (auto-updating)

install drops a hacknplan skill that teaches Claude how to drive these tools (id-resolution, feature breakdown, destructive-confirm rules, the portfolio / schedule / standup views). It's part of the npm package, so on every later server launch — which Claude runs via npx -y @mtuska/hacknplan-mcp (always the latest) — the server re-syncs the skill if the package version changed, keeping it current with no action from you.

The refresh only touches a skill copy this package installed (tracked by a .installed-by.json marker), so your own edits to other skills are never affected. Pass --skip-skill to register the MCP server without the skill.

Updating from a version before the skill existed (≤ 0.2.0): the auto-sync is update-only — it refreshes a skill it previously installed, but does not create one. If you installed the MCP server before the skill shipped (v0.3.0), re-run install once to add it:

npx -y @mtuska/hacknplan-mcp install --global

After that, every later upgrade auto-updates it on the next server launch.

Getting an API key

HacknPlan → your avatar → Settings → API → Create. Tick the scopes you need.

Install options

npx @mtuska/hacknplan-mcp install [options]

  -g, --global        Register for all projects (private, in ~/.claude.json). Default.
  -p, --project       Register for this repo only, shared via ./.mcp.json.
      --local         Register for this repo only, private to you.
      --api-key KEY   Provide the key non-interactively (else prompted / $HACKNPLAN_API_KEY).
      --name NAME     Server name to register (default: "hacknplan").
      --skip-skill    Register the MCP server only; don't install the Claude skill.
  -y, --yes           Non-interactive; use defaults and fail if the key is missing.

Non-interactive example (e.g. CI / dotfiles):

HACKNPLAN_API_KEY=hp_xxx npx @mtuska/hacknplan-mcp install --global -y

Project scope & secrets: --project writes the key into ./.mcp.json, which is typically committed. For shared repos prefer --global, or add .mcp.json to .gitignore.

The installer uses the official claude mcp add-json command when the Claude Code CLI is available, and otherwise writes the MCP config file directly.

Manual configuration

If you'd rather not use the installer, add this to your Claude MCP config (~/.claude.json for global, or a project .mcp.json):

{
  "mcpServers": {
    "hacknplan": {
      "command": "npx",
      "args": ["-y", "@mtuska/hacknplan-mcp"],
      "env": { "HACKNPLAN_API_KEY": "hp_your_key_here" }
    }
  }
}

Environment variables

Variable Required Purpose
HACKNPLAN_API_KEY yes Authorization: ApiKey <key> for the HacknPlan API.
HACKNPLAN_GROUPS no JSON map of group label → project names, for the portfolio grouping.
HACKNPLAN_GROUPS='{"Products":["Website","Mobile App"],"Ops":["Infra"]}'

Tools

75 tools across the HacknPlan surface. Highlights:

  • Read/introspecthacknplan_whoami, list_projects, get_project, list_work_items (rich server-side filters), get_work_item, list_stages/categories/tags/boards/milestones/importance_levels.
  • Search & focusfind_work_items (free-text + facet search to resolve a task by name into its id) and my_work (what's assigned to you, per-project or across all).
  • Work itemscreate_work_item (with story→task parent_id, assignees, dependencies, checklists), update_work_item (re-stage / re-prioritize / recategorize / re-estimate / reassign / move board·milestone), plan_feature (create a user story + child tasks + checklists in one call), delete_work_item, add_comment, sub-tasks, tags, user assignment.
  • Planningcreate_project/update_project/delete_project, stages, categories, tags, importance levels, boards (create_board/update_board/close_board/reopen_board), milestones (create_milestone/update_milestone/delete_milestone), dependencies.
  • Lifecycle — non-destructive archive/release via close_project/reopen_project and close_milestone/reopen_milestone.
  • Activity & metricsrecent_activity (the "what changed this week" / standup feed), log_work, list_work_logs, get_project_metrics, get_milestone_metrics, get_board_metrics (sprint burndown).
  • Design model*_design_element/*_design_element_type, plus get_design_element_metrics and design-element comments (list/add_design_element_comment) for the feature/knowledge tree.
  • Portfolio (cross-project)portfolio_overview (completion %, urgent / blocked / due-soon / overdue across all projects) and schedule_overview (a deadline countdown bucketed by horizon).

Destructive tools (delete_*, remove_dependency) require confirm=true.

Development

npm install
npm run build        # tsc -> dist/
npm run typecheck
npm run dev          # run the CLI from source via tsx
node dist/cli.js serve   # run the MCP server over stdio

Publishing

Releases are tag-driven via .github/workflows/release.yml. Pushing a vX.Y.Z tag validates that it matches package.json, publishes to npm, and creates the GitHub release:

npm version patch        # bumps package.json (CLI + server read it at runtime)
git push --follow-tags

Publishing uses npm Trusted Publishing (OIDC) with provenance — no NPM_TOKEN secret. One-time maintainer setup:

  1. Publish v0.1.0 once manually (npm publish --access public) to claim the @mtuska/hacknplan-mcp name.
  2. On npmjs.com → the package → Trusted Publishers, add this repo (mtuska/hacknplan-mcp) and the workflow filename (release.yml).

After that, every tag-driven release publishes with provenance and no long-lived token. Users can verify with npm audit signatures.

License

MIT © Montana Tuska

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured