AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
MCP verified MIT Self-run

Project Delivery Manager

mcp-yhooch-project-delivery-manager · by yhooch

SaaS-ready product delivery management platform with requirements, project context documents, tasks, bugs, workflows, realtime boards, and MCP-based AI agent access.

No reviews yet
0 installs
11 views
0.0% view→install

Install

$ agentstack add mcp-yhooch-project-delivery-manager

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No 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 Used
  • 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.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/mcp-yhooch-project-delivery-manager)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
1mo ago

Declared compatibility

Claude CodeClaude DesktopCursorWindsurf

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

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 →
Are you the author of Project Delivery Manager? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Project Delivery Manager

English | [简体中文](README.zh-CN.md)

Project Delivery Manager, or PDM, is a SaaS-ready product delivery management platform for small software teams. It brings requirements, intake, tasks, bugs, project context documents, custom workflows, version boards, realtime refresh, and MCP-based AI agent access into one TypeScript monorepo.

Keywords: product delivery management, project management, requirements management, document context library, issue tracking, bug tracking, workflow management, SaaS, MCP, OAuth, Next.js, NestJS, Prisma, PostgreSQL.

What It Solves

Small product teams often keep requirements, decisions, development tasks, test feedback, bug fixes, release scope, and delivery risks in separate tools or chat threads. PDM is designed to keep those objects connected and traceable without turning the workflow into a heavy enterprise process platform.

The core workflow is:

Register/login -> Organization -> Project space -> Version
  -> Requirement / Intake item -> Task / Bug
  -> Workflow action -> Timeline -> Workbench / Board / Exception view

Features

  • SaaS-ready organization and project-space isolation.
  • Username/password authentication with database sessions and HttpOnly cookies.
  • Organization members, project-space members, and role-based permissions.
  • Version management for delivery scope and progress tracking.
  • Requirement documents with Tiptap rich text, Markdown support, image uploads,

comments, attachments, and timeline history.

  • Space-level document context library for project knowledge capture, with

Markdown/plain text paste, Markdown/Word/HTML import, document list, search, filters, shared folders, detail reading, lightweight editing, re-import, archive/restore/delete, linked resources, tags, comments, attachments, timeline history, Mermaid diagram rendering, and fullscreen Mermaid preview.

  • Intake pool for collecting work sources and splitting accepted items into one

or more tasks.

  • Unified work item model for TASK and BUG, with separate user-facing pages.
  • Configurable lightweight workflow engine with states, actions, permissions,

action forms, and workflow versions.

  • My workbench, project overview, version board, and exception views.
  • Exception tracking for overdue, blocked, pending confirmation, pending

regression, and stale work items.

  • Tags for requirements, intake items, tasks, bugs, and documents.
  • Human-readable object codes such as REQ-12, INTAKE-8, TASK-42, and

BUG-17.

  • Realtime local refresh through SSE invalidation events.
  • Chinese and English UI with light, dark, and system themes.
  • MCP server integration for AI agents, protected by OAuth 2.1 + PKCE and scoped

Bearer tokens, including document and folder tools.

Screenshots

| My workbench | Version board | | --- | --- | | | |

| Workflow configuration | Task list | | --- | --- | | | |

| Document context library | | --- | | |

| MCP client authorizations | Organization settings | | --- | --- | | | |

| Account access | | --- | | |

MCP And AI Agent Access

PDM exposes selected product delivery capabilities through Model Context Protocol tools. The MCP surface includes context lookup, object code lookup, workbench, space overview, version board, exceptions, requirements, intake items, tasks, bugs, documents, shared document folders, comments, tags, and timelines.

Current tool names use the pdm.* namespace, for example:

  • pdm.context.get
  • pdm.object.lookup_code
  • pdm.requirement.create
  • pdm.work_item.execute_action
  • pdm.bug.create
  • pdm.document.create_from_markdown
  • pdm.document_folder.create
  • pdm.tag.replace_assignments

MCP requests do not reuse the web session cookie. They require OAuth access tokens scoped for the MCP resource, and business permissions are still enforced by the existing organization, project-space, object, role, and workflow rules.

Tech Stack

| Area | Technology | | --- | --- | | Monorepo | pnpm workspaces | | Web | Next.js, React, TypeScript | | API | NestJS, TypeScript | | Database | PostgreSQL, Prisma | | Shared contracts | Zod schemas in packages/shared | | UI | Tailwind CSS, Radix UI primitives, lucide-react | | Editor | Tiptap plus Markdown mode | | File storage | MinIO behind the API | | Realtime | Server-Sent Events | | E2E | Playwright | | Unit tests | Vitest |

Repository Layout

apps/
  api/        NestJS API application
  web/        Next.js web application
packages/
  shared/     Zod schemas, DTOs, enums, errors, OpenAPI helpers
prisma/       Prisma schema and migrations
tests/e2e/    Playwright API and UI E2E tests
deploy/       nginx production entrypoint config
docs/         README screenshot assets

Requirements

  • Node.js 22 or newer.
  • Corepack with pnpm 11.1.1.
  • PostgreSQL for the API database.
  • MinIO or another S3-compatible service for attachment flows.

For full local E2E, see [tests/e2e/README.md](tests/e2e/README.md).

Local Development

Enable the pinned package manager:

corepack enable
corepack prepare pnpm@11.1.1 --activate

Install dependencies:

corepack pnpm install

Prepare environment files:

cp .env.example .env

Generate Prisma client and run migrations as needed:

corepack pnpm --filter @project-delivery/api prisma:generate
corepack pnpm --filter @project-delivery/api prisma:migrate

Start both applications:

corepack pnpm dev

Or run them separately:

corepack pnpm dev:api
corepack pnpm dev:web

Verification

Run the standard local gate:

corepack pnpm verify

verify runs lint, typecheck, unit tests, and build.

Run the local E2E gate:

corepack pnpm verify:e2e:local

Run Docker-backed E2E:

corepack pnpm test:e2e:full

Useful targeted commands:

corepack pnpm lint
corepack pnpm typecheck
corepack pnpm test
corepack pnpm build
corepack pnpm test:e2e:list

Deployment

The repository includes a single-machine HTTP deployment path based on Docker Compose:

  • [docker-compose.prod.yml](docker-compose.prod.yml)
  • [docker-compose.prod.registry.yml](docker-compose.prod.registry.yml)
  • [Dockerfile.api](Dockerfile.api)
  • [Dockerfile.web](Dockerfile.web)
  • [deploy/nginx.conf](deploy/nginx.conf)
  • [.env.prod.example](.env.prod.example)

For HTTP deployment, SESSION_COOKIE_SECURE=false must match the actual origin. When moving to HTTPS, switch it back to secure cookie behavior. Public and private browser entry points are accepted when the request Host matches the browser Origin. OAuth/MCP metadata is derived from the current request Host, and attachment upload/download flows go through the API to the internal MinIO service. Production only needs the Nginx port exposed.

Current Status

The completed and accepted scope includes the core MVP workflow, TAG tagging, SEQ human-readable object codes, RT realtime local refresh, MCP AI agent system access, and the DOC document context library. Remaining release gates are full API/UI E2E against real dependencies, plus target-environment review for RT, MCP, and DOC behavior.

License

This project is licensed under the [MIT License](LICENSE).

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.