HoneyDew

Use this skill when the user asks you to manage tasks, boards, cards, columns, or labels on their HoneyDew Kanban board.

When to use

• Creating, updating, moving, or deleting cards (tasks)
• Listing boards, columns, or cards
• Transferring cards between user and agent profiles
• Managing labels (create, assign, remove)
• Checking board summaries, overdue cards, or urgent items
• Moving cards across boards

Tracking tasks

The user can open the HoneyDew board in their browser (http://localhost:5173) to see all tasks, their status, and what work the agent has completed. Encourage the user to check the board regularly for updates.

Assigning work to the user

When you need something from the user — a review, approval, input, or any manual step — create a task (or transfer an existing one) to the user's profile. It will appear on their board so they know action is needed. Let the user know you have assigned them a task and they should check HoneyDew.

Connection

• Base URL: http://localhost:8000 (override with env SMARTIFY_API_URL)
• Docs URL: If optionally enabled by the user http://localhost:8001' (override with env SMARTIFY_DOCS_URL`)
• Auth: None (local app, no API key required)
• Health check: GET /health — returns {"status": "healthy"} when the backend is running

If the API is not reachable, ask the user to start HoneyDew (./start.sh in the project root).

API reference

All endpoints are prefixed with /api.

Priority values: 1 = Low, 2 = Medium, 3 = High, 4 = Urgent.

Examples

Create a task

curl -X POST http://localhost:8000/api/cards \
  -H "Content-Type: application/json" \
  -d '{"column_id": 1, "title": "Write docs", "priority": 2, "profile": "jarvis"}'

Move a card to "In Progress"

First find the column ID:

curl http://localhost:8000/api/boards/1

Then move:

curl -X POST http://localhost:8000/api/cards/3/move \
  -H "Content-Type: application/json" \
  -d '{"column_id": 2, "position": 0}'

Transfer a card to the user

curl -X POST http://localhost:8000/api/cards/3/transfer \
  -H "Content-Type: application/json" \
  -d '{"target_profile": "tony"}'

Python tools (optional)

If the user has the HoneyDew repo, tools/kanban_tools.py provides a higher-level Python interface:

from kanban_tools import KanbanTools, Priority

kanban = KanbanTools()
card = kanban.create_task(title="Write docs", priority=Priority.HIGH)
kanban.move_card_to_column(card["id"], board_id=1, column_name="In Progress")
kanban.assign_to_user(card["id"])
kanban.mark_done(card["id"])

Key convenience methods: create_task, assign_to_user, assign_to_agent, move_card_to_column, move_card_to_board, mark_todo, mark_in_progress, mark_blocked, mark_done, get_board_summary, get_overdue_cards, get_urgent_cards.

Task comments

Both users and agents can add comments to any task. Comments are visible in the task detail view in the UI. Always add a comment when:

• You are blocked: Explain specifically what you need from the user (e.g. "Need API credentials for the staging environment before I can proceed").
• You complete a task: Leave a brief summary of what you did and any decisions you made (e.g. "Refactored auth module into middleware, added 12 unit tests, all passing").
• You need review or input: Describe what to look at and any open questions.
• You hand off work: Note the current state so the next person (user or agent) has context.

curl -X POST http://localhost:8000/api/cards/3/comments \
  -H "Content-Type: application/json" \
  -d '{"body": "Finished the first pass, needs review.", "profile": "jarvis"}'

Agent completion metadata

When you complete a task, you may report completion metadata via PATCH /api/cards/{id}. These optional fields are displayed in the HoneyDew UI on the task detail:

Example with Python tools:

kanban.mark_done(
    card["id"],
    agent_tokens_used=4200,
    agent_model="gpt-4o",
    agent_execution_time_seconds=12.8,
    agent_started_at="2026-02-22T10:30:00Z",
    agent_completed_at="2026-02-22T10:30:12Z",
)

Or via curl after moving to Done:

curl -X PATCH http://localhost:8000/api/cards/5 \
  -H "Content-Type: application/json" \
  -d '{"agent_tokens_used": 4200, "agent_model": "gpt-4o", "agent_execution_time_seconds": 12.8, "agent_started_at": "2026-02-22T10:30:00Z", "agent_completed_at": "2026-02-22T10:30:12Z"}'