Install
$ agentstack add mcp-martinkuhne-devstack ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
devstack
Mission
devstack is a project management MCP server and UI to enable better interaction with code agents. It provides a Project → Deliverable → AgentTask hierarchy to separate planning from implementation. If you are cost conscious (or otherwise token constrained), use a frontier model for planning and an efficient model for coding.
Deliverables and tasks are tracked through their lifecycle (Draft, Ready, Implement, Review, Done) but the system is not opinionated on state transitions.
There is a fully fledged admin UI to add features, or review/approve AI generated tasks:
Features
MCP Server
Provides tools for AI coding agents (OpenCode, Claude, etc.) to interact with DevStack data via the Model Context Protocol:
- Read, create, and update Projects, Deliverables, and AgentTasks
- Smart "get next" tools that find the next item to work on based on priority
- HTTP streaming protocol, JSON-RPC 2.0, health check endpoint
MCP Tools
| Tool | Description | |------|-------------| | get_projects | Read all projects from DevStack | | get_project | Read a project by its ID | | create_project | Create a new project in DevStack | | get_deliverable | Read a deliverable by its ID | | get_next_deliverable | Find the next deliverable in Implement status for a project | | create_deliverable | Create a new deliverable (Feature) in DevStack | | update_deliverable | Modify an existing deliverable in DevStack | | update_deliverable_status | Change the state of a deliverable in DevStack | | get_task | Read an agent task by its ID | | get_next_task | Find the next task to work on for a project | | create_task | Create a new agent task in DevStack | | update_task | Modify an existing agent task in DevStack | | update_task_status | Change the state of an agent task in DevStack |
Admin UI
A web interface built with React, Tailwind, and ShadCN for human interaction:
- CRUD management for all entities (Projects, LLMs, Deliverables, AgentTasks)
- Dashboard showing deliverable counts by status
- Markdown rendering for rich text fields
- Search and navigation by project
- Status transitions via dropdown
GraphQL API
Backend data layer with full CRUD over all entities, filtering, sorting, and paging:
- Projects - top-level containers with repository tracking
- Deliverables (Features, Spikes, Defects, Maintenance) - full lifecycle with design, acceptance criteria, execution plan, security/performance impact, test and deployment plans
- AgentTasks - granular work items linked to deliverables
- LargeLanguageModels - configurable model definitions with cost and complexity tracking
Note
This was intentionally intended to be a fully automated coding system. For the time being, I am focusing on the mcp server and UI parts as the most valuable aspects.
References and inspiration
Source & license
This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: MartinKuhne
- Source: MartinKuhne/devstack
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.