MCP Native

Your AI Teammate, Not Just a Chatbot

Dragon Planner is built on the Model Context Protocol from the ground up. Connect any MCP-compatible AI assistant directly to your project data — with real auth, real permissions, and a full audit trail.

Board view showing work items managed by AI across status columns

33 Tools Across 9 Categories

Your AI assistant gets direct access to your project management workflow. No fragile integrations, no screen scraping — structured tool calls with full type safety.

  • Work items — create, update, move, delete, search, and complete items across a 7-level hierarchy
  • Smart recommendations — ask "what should I work on next?" and get dependency-aware suggestions
  • Sprint & PI planning — check sprint status, get summaries, assign items, review program increments
  • Time tracking, dashboards, team info, attachments, comments, and Jira sync — all from your terminal
claude — reading AI context

you: Pick up DP-36 and start working on it.

→ get_work_item DP-36

DP-36 "Invoice PDF generation" [P1]

AI Context:

Use the billing API to pull invoice line items.

Generate PDF with the company branding template

in /templates/invoice.html.

Include tax calculation from TaxService.

Follow the existing EmailService pattern for

delivery — Channel → BackgroundService → send.

I'll follow the instructions — pulling from the billing API, using the HTML template with tax calculation, and wiring delivery through the existing email channel pattern.

Per-Item AI Context

Every work item can carry rich text AI context — guidance that tells your AI assistant how to handle that specific piece of work. Context travels with the item, not lost in chat history.

  • Rich text blocks stored directly on each work item
  • AI reads instructions before taking action — no guessing
  • Teams can encode domain knowledge, coding standards, or acceptance criteria
  • Instructions persist across sessions and assistants
claude — what's next?

you: What should I work on next?

→ get_next_recommendation limit=5

Top items ready to start:

[P1] DP-38 Multi-tenant Data Export 13 SP

[P2] DP-40 Evaluate real-time collaboration 5 SP

[P3] DP-39 Dark mode toggle 3 SP

3 items skipped — blocked by unresolved dependencies.

you: Start the spike. Assign it to me.

→ update_work_item DP-40 status=InProgress assigneeId=...

DP-40 is now In Progress and assigned to you.

Smart Recommendations Engine

No more digging through a stale backlog to figure out what to do next. The recommendation engine understands your project's dependency graph and surfaces the highest-impact work.

  • Dependency-aware — only recommends items whose blockers are all resolved
  • Priority-sorted — highest priority items surface first
  • Sprint-scoped — recommendations respect your current sprint context
  • Works for both human users and AI assistants
Audit log showing team activity with source attribution for web and MCP actions

Full Audit Trail

Every action taken by an AI assistant is logged with the same fidelity as human actions. You always know who did what, when, and whether it came from the web app or an AI tool.

  • Source attribution — every log entry marks whether it came from web or MCP
  • User-level tracing — actions tied to the authenticated user, not a generic bot account
  • Immutable history — audit logs cannot be edited or deleted
  • Visible in dashboards and activity feeds alongside human actions

33 tools. One conversation.

Real examples of Claude Code working with Dragon Planner through MCP. Every tool call below is a real capability — not a mockup.

Morning Standup

claude — sprint check-in

you: What's the status of our current sprint?

→ get_current_sprint

Sprint 7 · "Mobile beta prep and design tokens"

Apr 13 – Apr 24 · Active · 45 SP committed · 160h capacity

you: How are we tracking? Show me the dashboard.

→ get_dashboard

Here's Sprint 7 at a glance:

5
Todo
10
In Progress
21
Done
2
Blocked

2 blocked items need attention — DP-29 and DP-36.

you: What's blocking DP-29?

→ get_work_item DP-29

Bug DP-29 [P1] "Webhook retry storms on 502 errors"

Assigned to Marcus Johnson · Sprint 7 · 3 SP

Payment webhook retries cause duplicate processing when backend returns 502

you: Add a comment that I'm bumping the proxy buffer size to fix the 502s.

→ add_comment DP-29

Comment added to DP-29: "Bumping proxy_buffer_size to 16k — that should resolve the 502s on large webhook payloads."

Building Features

claude — feature implementation

you: I need to build the invoice PDF feature. Set my context to Dragon Platform and show me what's there.

→ set_context project="DP"

Context set to Dragon Platform in Engineering workspace.

→ get_work_item DP-36

Story DP-36 [P1] "Invoice PDF generation"

Status: Blocked · Sprint 7 · 5 SP · Parent: DP-22 (Billing)

AI Context: "Use the billing API to pull invoice line items. Generate PDF with the company branding template in /templates/invoice.html. Include tax calculation."

you: Unblock it and mark it in-progress. Create three subtasks: pull billing data, render PDF template, email delivery.

→ update_work_item DP-36 status=InProgress

→ create_work_item SubTask parent=DP-36 ×3

DP-36 is now In Progress.

Created 3 subtasks:

• DP-48 "Pull invoice line items from billing API"

• DP-49 "Render PDF from HTML template with branding"

• DP-50 "Send invoice email via email service"

you: Assign the billing task to me and log 2 hours on it.

→ update_work_item DP-48 assigneeId=...

→ log_time DP-48 hours=2

DP-48 assigned to you. Logged 2.0h for today. Total: 2.0h / no estimate set.

Sprint Planning

claude — sprint planning session

you: We're planning Sprint 8. What's the team's recent velocity?

→ get_sprint_summary sprintId=sprint6

→ get_sprint_summary sprintId=sprint7

Last two sprints:

• Sprint 6: 16 SP committed, 16 SP completed (100%)

• Sprint 7: 45 SP committed, 24 SP completed so far (53% with 5 days left)

Average velocity: ~20 SP. Sprint 7 may have been overcommitted.

you: What should we pull into Sprint 8? Show me recommendations.

→ get_next_recommendation limit=8

Top 5 ready items (dependencies resolved, priority-sorted):

[P1] DP-38 "Multi-tenant Data Export" — Epic, 13 SP

[P2] DP-44 "Jira Integration" — Epic, 21 SP

[P2] DP-40 "Evaluate real-time collaboration" — Spike, 5 SP

[P3] DP-39 "Dark mode toggle" — Story, 3 SP

[P3] DP-41 "Bulk import from CSV" — Story, 5 SP

Based on ~20 SP velocity, I'd suggest the spike + dark mode + bulk import = 13 SP, leaving room for carryover.

you: Good call. Assign those three to Sprint 8.

→ assign_to_sprint DP-40,DP-39,DP-41 sprintId=sprint8

3 items assigned to Sprint 8. Total: 13 SP committed.

Search & Jira Sync

claude — cross-project search

you: Search across all projects for anything related to "rate limiting"

→ search_work_items query="rate limiting"

Found 4 work items across 2 projects:

• API-2 "Rate Limiting & Throttling" — Epic, Done (API Gateway)

• API-3 "Implement sliding window rate limiter" — Story, Done

• API-4 "Per-tenant rate limit configuration" — Story, Done

• API-10 "Rate limiter leaks memory under high concurrency" — Bug, Done

you: Generate a Jira payload for API-2 so I can push it to our corporate Jira.

→ generate_jira_payload API-2

{
  "fields": {
    "summary": "Rate Limiting & Throttling",
    "issuetype": { "name": "Epic" },
    "priority": { "name": "Critical" },
    "description": "Token bucket rate limiter with per-tenant and per-endpoint limits",
    "labels": ["api", "rate-limit"]
  }
}

Copy this payload to your Jira MCP server, then use update_jira_reference to link them.

End of Day

claude — wrapping up

you: I finished the CSV parser subtask and the billing data pull. Mark both done and log my time.

→ mark_complete DP-48

→ mark_complete DP-49

→ log_time DP-48 hours=3 note="Billing API integration"

→ log_time DP-49 hours=4.5 note="PDF template rendering"

Marked DP-48 and DP-49 as Done.

Logged 3.0h on DP-48, 4.5h on DP-49. Total today: 9.5h.

you: How many items are left on DP-36?

→ get_work_item DP-36 includeChildren=true

DP-36 "Invoice PDF generation" — 3 subtasks:

✓ DP-48 "Pull invoice line items from billing API"

✓ DP-49 "Render PDF from HTML template with branding"

• DP-50 "Send invoice email via email service" — Todo

1 remaining. Finish DP-50 and you can cascade-complete the parent.

you: Almost there. See you tomorrow.

Good day. 2 items completed, 9.5h logged. DP-36 is 67% done. 🤠

Context 5Projects 2Work Items 12Time Logs 3Sprints 3PIs 2Dashboard 1Team 1Jira 4

33 tools available via Model Context Protocol

Every Tool, At a Glance

The full MCP tool surface your AI assistant can use. Each tool is a structured API call — typed inputs, typed outputs, no guesswork.

Context (5)

  • list_workspaces

    List all workspaces you have access to

  • create_workspace

    Create a new workspace

  • set_context

    Set active workspace and/or project by name, prefix, or UUID

  • get_context

    Show current active workspace and project

  • clear_context

    Clear active project or workspace context

Projects (2)

  • list_projects

    List projects in a workspace

  • create_project

    Create a new project in a workspace

Work Items (12)

  • list_work_items

    List work items with optional filters

  • get_work_item

    Get full details including direct children

  • create_work_item

    Create with hierarchical validation

  • update_work_item

    Partial update — only provided fields change

  • mark_complete

    Mark done, optionally cascade to descendants

  • delete_work_item

    Delete item and all descendants

  • move_work_item

    Move to a different parent

  • upload_attachment

    Attach files via base64 or curl upload

  • get_next_recommendation

    Priority + dependency-aware next-item suggestions

  • search_work_items

    Search items across all projects

  • add_comment

    Add a comment to a work item

  • assign_to_sprint

    Assign or unassign items to a sprint

Time Logs (3)

  • list_time_logs

    List entries with hours, dates, and totals

  • log_time

    Log time spent on a work item

  • delete_time_log

    Delete a time log entry

Sprints (3)

  • get_current_sprint

    Get the currently active sprint

  • get_sprint

    Get sprint details by ID

  • get_sprint_summary

    Completion counts, points, and incomplete items

Program Increments (2)

  • list_program_increments

    List PIs, filter by year or quarter

  • get_program_increment

    Full PI details with all sprints and dates

Dashboard (1)

  • get_dashboard

    Status, type, and priority counts, story points, blocked and recently completed items

Team (1)

  • list_team_members

    List organization members with roles and seat types

Jira Integration (4)

  • list_unsynced_items

    Items without a Jira key — export candidates

  • generate_jira_payload

    Generate Jira-formatted payload for export

  • update_jira_reference

    Store a Jira key on a work item

  • bulk_update_jira_references

    Update Jira keys for multiple items

Enterprise Authentication

OAuth 2.0 with PKCE flow. Fine-grained permissions per resource type — read, write, and delete scoped independently.

Builder vs Collaborator Seats

Builder seats unlock AI capabilities via MCP. Collaborator seats are for web-only team members — pay only for what you use.

Plan-Aware Rate Limiting

Each plan tier has its own rate limits and connection caps. Scale your AI usage as your team grows.

Explore more features

Dashboards

Three levels of dashboards — organization, workspace, and project — so you always know where things stand.

Work Items

Seven work item types with configurable hierarchy rules — strict or relaxed. Dependencies and cascade completion.

Sprints

Program increments, sprint lifecycles, boards, and a recommendation engine that surfaces what to ship next.

Collaboration

Comments, real-time notifications, role-based access, and file attachments — everything your team needs to stay in sync.

Ready to plan like a dragon?

Free to start. No credit card required. Upgrade when your team grows.

Get started free→