Skip to main content
Glama
scrumdo-mcp

scrumdo-mcp

Official
by ScrumDoLLC

scrumdo-mcp

Connect any MCP-compatible AI tool (Claude Code, Cursor, Windsurf, and others) directly to your ScrumDo boards.

Once installed, your AI assistant can read cards, move them, create tasks, post comments, and search across your board — without you copy-pasting anything.


Installation

pip install scrumdo-mcp

Related MCP server: kanban-mcp

Cursor quick-start

1 — Install

# To run the server only:
pip install scrumdo-mcp

# To run tests too:
git clone https://github.com/ScrumDoLLC/scrumdo-mcp.git
cd scrumdo-mcp
pip install -e ".[dev]"
pytest tests/ -v

2 — Get your token

Log in to ScrumDo → your org → Settings → API Tokens → Create Token. Copy it — shown once only.

3 — Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "scrumdo": {
      "command": "scrumdo-mcp",
      "env": {
        "SCRUMDO_TOKEN": "your-token-here",
        "SCRUMDO_ORG": "your-org-slug",
        "SCRUMDO_PROJECT": "your-default-project-slug"
      }
    }
  }
}

Your org and project slugs are the short names in your board URL: app.scrumdo.com/my-company/engineering

4 — Restart Cursor

Done. In any Cursor chat you can now ask:

  • "What cards are in the current sprint?"

  • "Move ENG-42 to In Review"

  • "Add a comment to ENG-42: PR is up for review"

  • "List all cards assigned to me"


Setup

Step 1 — Get your token

Log in to ScrumDo → your organization → Settings → API Tokens → Create Token.

Copy the token — it is only shown once. This is your personal key; keep it private.

Step 2 — Configure your AI tool

Find your tool's MCP config file and add the scrumdo server entry:

Tool

Config file

Claude Code

~/.claude.json (home dir — not ~/.claude/claude.json)

Claude Desktop

~/Library/Application Support/Claude/claude_desktop_config.json (macOS) · %APPDATA%\Claude\claude_desktop_config.json (Windows)

Cursor

~/.cursor/mcp.json

Windsurf

~/.codeium/windsurf/mcp_config.json

Claude Code — the reliable way. Rather than hand-editing the JSON (it's easy to edit the wrong ~/.claude/claude.json by mistake), let the CLI write it:

claude mcp add scrumdo \
  -e SCRUMDO_TOKEN=your-token-here \
  -e SCRUMDO_ORG=your-org-slug \
  -e SCRUMDO_PROJECT=your-project-slug \
  -- scrumdo-mcp
{
  "mcpServers": {
    "scrumdo": {
      "command": "scrumdo-mcp",
      "env": {
        "SCRUMDO_TOKEN": "your-token-here",
        "SCRUMDO_ORG": "your-org-slug",
        "SCRUMDO_PROJECT": "your-default-project-slug"
      }
    }
  }
}

Your org slug and project slug are the short names in your board URL: app.scrumdo.com/my-company/engineering

Step 3 — Restart your AI tool

Done. Your AI assistant now has direct access to your board.


What you can do

Once connected, just talk to your AI tool naturally:

"What's the status of ENG-42?"
"Move ENG-42 to In Review and add a comment saying the PR is up"
"List all cards assigned to me in the current sprint"
"Create a sub-task on ENG-42: write release notes"
"Search for cards about the login bug"
"What did the team work on this week?"
"Block ENG-42 — waiting on design approval"
"Move ENG-42 to the Sprint 14 iteration"
"Set the due date on ENG-42 to 2026-04-30"
"Assign ENG-42 to Sarah"

Available tools (132 total)

Group

Tools

Boards

list_boards, get_board, get_board_cells, list_iterations, list_milestones, list_labels, list_epics

Cards

list_cards, get_card, card_schema, find_card, create_card, update_card, move_card, move_card_to_iteration, set_card_field, set_card_fields, archive_card, assign_card, add_card_label, remove_card_label

Blockers

list_blockers, block_card, unblock_card

Tasks

list_tasks, create_task, complete_task, reopen_task, update_task, delete_task

Comments

list_comments, add_comment, delete_comment

Attachments

add_attachment

Fields

list_custom_fields, get_card_field, get_all_card_fields

Members

list_members, find_member

Search

search_cards, search_by_field_value

Activity

log_activity, get_activity_log, get_workspace_activity

Webhooks

list_webhooks, create_webhook, delete_webhook

Time

list_time_entries, log_time

Spec

get_card_spec, set_card_spec, patch_card_spec, get_spec_history, list_card_spec_documents, set_card_spec_document, restore_spec_version — get/set/patch operate on the primary (requirements) doc; the multi-doc trio lists per-doc_type documents, writes a specific one (human-only), and restores an accepted version forward

Spec proposals

generate_spec_proposal, list_spec_proposals, accept_spec_proposal, reject_spec_proposal, request_spec_proposal_changes, revise_spec_proposal, get_decision_inbox, read_spec_proposal, preview_spec_decision, attest_spec_understood — all human-only; deciding from an MCP session is gated: get_decision_inbox → read_spec_proposal (delivers the content; the delivery IS the evidence, so it records the evidence-open + viewed the gate requires) → attest_spec_understood → preview_spec_decision mints a confirm_token (bound to the proposal's current version, 10-min TTL) → pass it to accept/reject/request_changes. When the org has governed_review_sessions_enabled, the read + attestation are required — accept fails with needs_understood without them

GitHub

get_github_repos, list_card_github_links, link_github_pr, link_github_commit, link_github_issue

Cockpit commands

cockpit_help (network-free list of every governed command + the MCP tool that runs it — ask "what can I do?"), invoke_cockpit_command (governed dispatcher for ANY catalog command by id — executes loop.status/pause/resume + skill.*, governance-validates the rest), research_card, run_card_tests, tasks_from_spec, get_card_memory, clear_card_memory — every command in get_effective_governance's catalog is reachable

Cockpit

get_card_cockpit_context, get_effective_governance, get_mcp_capabilities, send_cockpit_chat, draft_spec_from_card — the Card AI Cockpit bridge: one-call card context, the governed command policy for a card, this bridge's own tool/connection surface, plus human-only cockpit writes (chat a board agent, draft a spec doc). MCP writes are attributed to the cockpit timeline via X-Spryng-Source: mcp

Agents

get_agent_identity, list_agent_accounts

Agent runs

start_agent_run, get_agent_run, list_agent_runs, approve_agent_plan, accept_proof, request_agent_replan, execute_task, report_agent_progress, cancel_agent_run — approve_agent_plan / accept_proof / request_agent_replan / execute_task are human-only (run as a human principal); execute_task runs a spec-derived task with an agent (Todo→Doing→Reviewing)

Loops & verification

start_loop, start_verification_loop, get_loop_status, list_active_loops, pause_loop, resume_loop, cancel_loop, get_loop_state, update_loop_state, get_verification_status, run_verifier, verify_card, log_loop_step, attach_evidence, route_to_agent, list_skills, load_skill

Intelligence

get_velocity_forecast, get_spec_complexity, check_spec_drift, verify_behavior_contract

Shared memory

get_handoff_brief (call FIRST when picking a card up — "since you last touched this"), read_blackboard, post_blackboard_note, drop_blackboard_note (the card's live working notes; agents may post), promote_blackboard_note, add_card_memory (promote a note / add durable saved context — human-only), get_room_context, add_room_context, curate_room_context, run_distiller (the room/board standing-rules library; add/curate/distill are room-manager only), list_memory_disputes, resolve_memory_dispute (contradicting saved entries; resolving is human-only)

Notifications

list_notifications, notification_counts, mark_notification (read / acknowledge), mark_all_notifications_read, wait_for_notifications (long-poll push channel — blocks until something arrives or timeout_s elapses, so agents can wait on a decision instead of spinning)

For a governed verification loop, an orchestrator agent calls start_verification_loop (by VerificationProfile slug or inline proof_requirements/verifier_agent), the Maker (Grok/Codex) implements and calls run_verifier against the accepted spec (never self-verifies), and log_loop_step / attach_evidence write the audit trail to the card. When an agent runs inside a loop (SPRYNG_LOOP_ID set), the loop-scoped tools default to that loop, so loop_id is optional.


Environment variables

Variable

Default

Description

SCRUMDO_TOKEN

—

Required. API token from Settings → API Tokens (or an agent's token, for AI Agent runs)

SCRUMDO_BASE_URL

https://app.spryng.io

API base URL

SCRUMDO_ORG

—

Your organization slug

SCRUMDO_PROJECT

—

Default project slug

SCRUMDO_AGENT_RUN_ID

—

Optional. AI Agent run id this MCP is driving. When set, every write sends the X-Spryng-Agent-Run header so the run's audit trail attributes the write (change_source='agent_run'). Requires SCRUMDO_TOKEN to be that agent's own token, and the run to belong to it.

SCRUMDO_CLIENT_NAME

mcp

Optional. The host tool driving this bridge (codex / claude-code / cursor). Sent as X-Spryng-Client alongside X-Spryng-Source: mcp so the Card AI Cockpit timeline can show "via MCP ()". (SPRYNG_CLIENT_NAME accepted as an alias; SCRUMDO_CLIENT_VERSION optionally adds a version.)

SPRYNG_LOOP_ID

—

Optional. The governed loop this MCP is running inside. When set, writes carry the X-Spryng-Loop header (attributed to the loop's timeline) and the loop-scoped tools (log_loop_step, attach_evidence, get_verification_status) default their loop_id to it — so in-loop agents call them without an id. (SCRUMDO_LOOP_ID is accepted as an alias.)


Token scope

Your API token is restricted to your organization's board data only — cards, tasks, comments, members, iterations. It cannot access billing, account settings, or any other organization's data. Revoke it at any time from Settings → API Tokens.


What is MCP?

Model Context Protocol is an open standard for connecting AI tools to external services. Claude Code, Cursor, Windsurf, and other AI editors support it natively. Install the server once; any MCP-compatible tool can use it.


License

MIT

Available Tools

45 tools
add_card_labelA

Add a single label to a card, preserving existing labels.

Args: card_ref: Card reference, e.g. 'ON-914'. label_id: Label id to add. Use list_labels() to find ids.

Inputs: card_ref (required), label_id (required).

add_commentA

Post a comment on a card. Markdown is supported.

Args: card_ref: Card reference, e.g. 'ON-914'. body: Comment text (markdown). Code blocks, lists, and headers all render in the ScrumDo UI.

Returns the created comment object with id and timestamp.

Inputs: card_ref (required), body (required).

archive_cardB

Archive a card by moving it to the project's Archive iteration. The card is hidden from the board but recoverable via the archive view.

Args: card_ref: Card reference, e.g. 'ON-914'.

Inputs: card_ref (required).

assign_cardA

Set the assignees on a card, replacing any current assignees.

Args: card_ref: Card reference, e.g. 'ON-914'. assignee_ids: List of user IDs. Pass an empty list to unassign all.

Inputs: card_ref (required), assignee_ids (required).

block_cardA

Mark a card as blocked with a reason.

Creates a blocker record on the card. Multiple blockers can exist simultaneously; the card remains blocked until all are resolved.

Args: card_ref: Card reference, e.g. 'ON-914'. reason: Human-readable explanation of what is blocking the card. is_show_stopper: Flag this as a show-stopper blocker (default False). external: Flag as an external dependency (default False).

Returns the created blocker object including its id (needed to unblock).

Inputs: card_ref (required), reason (required), is_show_stopper, external.

complete_taskB

Mark a task as complete.

Args: card_ref: Card reference, e.g. 'ON-914'. task_id: Task id (from list_tasks).

Inputs: card_ref (required), task_id (required).

create_cardB

Create a new card on the board.

Args: summary: Card title / summary (required). description: Rich text description (markdown supported). points: Story points / estimate. due_date: Due date in YYYY-MM-DD format. cell_id: Board column to place the card in. iteration_id: Iteration/sprint to assign the card to. assignee_ids: List of user IDs to assign. Use list_members() to find ids. label_ids: List of label IDs to attach. extra_fields: Custom field values as {field_id: value} dict.

Returns the full created card object including its assigned reference (e.g. 'ON-915').

Inputs: summary (required), description, points, due_date, +5 more.

create_taskA

Create a new task on a card.

Args: card_ref: Card reference, e.g. 'ON-914'. description: Task description (plain text or markdown). assignee_id: Optional user ID to assign the task to.

Returns the created task object.

Inputs: card_ref (required), description (required), assignee_id.

create_webhookA

Create a new webhook on the project.

Args: url: HTTPS endpoint that will receive webhook POST requests. events: List of event names to subscribe to, e.g. ['story.created', 'story.updated', 'story.moved', 'task.completed', 'comment.created']. secret: Optional HMAC secret for payload signature verification.

Returns the created webhook object with its id.

Inputs: url (required), events (required), secret.

delete_commentA

Delete a comment from a card.

Args: card_ref: Card reference, e.g. 'ON-914'. comment_id: Numeric comment id from list_comments().

Inputs: card_ref (required), comment_id (required). Returns: result

delete_taskA

Delete a task from a card.

Args: card_ref: Card reference, e.g. 'ON-914'. task_id: Task id (from list_tasks).

Inputs: card_ref (required), task_id (required). Returns: result

delete_webhookA

Delete a webhook by id.

Args: webhook_id: Webhook id from list_webhooks().

Inputs: webhook_id (required). Returns: result

find_cardA

Find a card by its reference ID (e.g. 'ON-914'). Alias for get_card — useful when you only know the ID string.

Args: card_ref: Card reference like 'ON-914'.

Inputs: card_ref (required).

find_memberA

Find a member by name, username, or email.

Args: query: Name, username, or email fragment to search.

Returns matching members from the org member list.

Inputs: query (required). Returns: result

get_activity_logA

Get the structured activity log for a card, with optional filters.

Returns only entries written via log_activity() — plain comments are excluded. Results are newest-first.

Args: card_ref: Card reference, e.g. 'ON-914'. agent: Filter by agent name (substring, case-insensitive). user: Filter by user (substring, case-insensitive). action: Filter by action label (substring, case-insensitive). environment: Filter by environment name (exact). milestone: Filter by milestone tag (exact). limit: Maximum entries to return (default 50).

Inputs: card_ref (required), agent, user, action, +3 more. Returns: result

get_all_card_fieldsA

Get all custom field values currently set on a card.

Returns the full list of custom fields with their names and current values.

Args: card_ref: Card reference, e.g. 'ON-914'.

Inputs: card_ref (required). Returns: result

get_boardA

Get a board (project) with its cells (columns), custom field definitions, and key statistics.

Args: project_slug: Board slug to fetch. Defaults to the configured project.

Inputs: project_slug.

get_board_cellsA

List the columns (cells) of a board.

Each cell has an id, name, and position. Use the id when moving cards.

Args: project_slug: Board slug. Defaults to the configured project.

Inputs: project_slug. Returns: result

get_cardA

Get a card's full detail including all custom fields, task list, comment count, labels, assignees, and linked cards.

Args: card_ref: The card reference, e.g. 'ON-914' or 'Q1-42'.

Inputs: card_ref (required).

get_card_fieldA

Get the current value of a single custom field on a card.

Args: card_ref: Card reference, e.g. 'ON-914'. field_id: Numeric custom field id from list_custom_fields().

Returns {'field_id': …, 'name': …, 'value': …} or {'field_id': …, 'value': null} if the field is not set.

Inputs: card_ref (required), field_id (required).

get_workspace_activityA

Get the activity log across ALL cards in the workspace.

Aggregates log_activity() entries from every card, sorted newest-first. Filterable by agent, user, action type, environment, and milestone.

This is the workspace-level view requested for board observability — answers questions like: • "What did claude-code do today?" • "What was deployed to stage this week?" • "Show me all activity on the book-kit milestone"

Args: action: Filter by action label (substring). agent: Filter by agent name (substring). user: Filter by user (substring). environment: Filter by environment (exact). milestone: Filter by milestone tag (exact). limit: Max entries across all cards (default 100).

NOTE: This scans recent cards; for large boards use per-card get_activity_log() for targeted queries.

Inputs: action, agent, user, environment, +2 more. Returns: result

list_blockersA

List all active (unresolved) blockers on a card.

Each blocker includes id, reason, blocked_date, blocker (user), is_show_stopper, and external flags.

Args: card_ref: Card reference, e.g. 'ON-914'.

Inputs: card_ref (required). Returns: result

list_boardsA

List all boards (projects) in the organisation.

Returns each board's slug, name, description, and member count. Use get_board() for full detail on a specific board.

Inputs: no named parameters. Returns: result

list_cardsA

List cards on the board with optional filters.

Args: cell_id: Filter by board column (cell) id. iteration_id: Filter by iteration/sprint id. assignee: Filter by assignee username or email. label: Filter by label name. status: Filter by status string. page: Page number (default 1). limit: Results per page (default 25, max 100).

Returns a dict with count, next/previous page URLs, and items list. Each item includes id, local_id, summary, points, cell, assignees, labels, extra_fields (custom fields), and created/updated timestamps.

Inputs: cell_id, iteration_id, assignee, label, +3 more.

list_commentsA

List all comments on a card, newest first.

Args: card_ref: Card reference, e.g. 'ON-914'. The card's numeric id is resolved automatically.

Returns a list of comments with id, author, body (markdown), and created timestamp.

Inputs: card_ref (required). Returns: result

list_custom_fieldsA

List all custom field definitions for the board.

Each field has an id, name, field_type, and optional choices. Use the id when calling set_card_field() or update_card(extra_fields=…).

Args: project_slug: Board slug. Defaults to the configured project.

Inputs: project_slug. Returns: result

list_epicsB

List all epics in the project.

Args: project_slug: Board slug. Defaults to the configured project.

Inputs: project_slug. Returns: result

list_iterationsB

List all iterations (sprints / milestones) in the project.

Args: project_slug: Board slug. Defaults to the configured project.

Inputs: project_slug. Returns: result

list_labelsA

List all labels defined in the project.

Args: project_slug: Board slug. Defaults to the configured project.

Inputs: project_slug. Returns: result

list_membersA

List members of the organisation or a specific project.

Each member has an id, username, email, display_name, and role.

Args: project_slug: If provided, returns project-level members only. If omitted, returns all org members.

Inputs: project_slug. Returns: result

list_tasksA

List all tasks (checklist items) on a card.

Args: card_ref: Card reference, e.g. 'ON-914'.

Returns a list of tasks, each with id, description, complete flag, assignee, and position.

Inputs: card_ref (required). Returns: result

list_time_entriesA

List time entries, optionally filtered to a specific card.

Args: card_ref: Card reference, e.g. 'ON-914'. If omitted, returns all project time entries.

Inputs: card_ref. Returns: result

list_webhooksA

List all webhooks configured on the project.

Returns each webhook's id, url, events, and active state.

Inputs: no named parameters. Returns: result

log_activityA

Write a structured activity log entry to a card.

Entries are stored as machine-parseable comments and rendered in the ScrumDo UI as human-readable updates. They are filterable via get_activity_log().

Args: card_ref: Card reference, e.g. 'ON-914'. action: Short action label, e.g. 'started', 'deployed', 'reviewed', 'snapshot', 'seeded', 'PR opened'. detail: Longer description or context (markdown). agent: Agent/bot name writing the log (e.g. 'claude-code', 'codex', 'github-actions'). user: Human user involved (username or email). environment: Environment name (stage, integration, branch). milestone: Milestone or release tag, if applicable. task_ref: Related task or sub-item reference. extra: Any additional structured data to attach.

Returns the created comment object.

Inputs: card_ref (required), action (required), detail, agent, +5 more.

log_timeA

Log time spent on a card.

Args: card_ref: Card reference, e.g. 'ON-914'. minutes: Time spent in minutes. description: What was done during this time. date: ISO date string (YYYY-MM-DD). Defaults to today. user_id: User to log time for. Defaults to the token owner.

Returns the created time entry.

Inputs: card_ref (required), minutes (required), description, date, +1 more.

move_cardA

Move a card to a different column (cell) on the board.

Args: card_ref: Card reference, e.g. 'ON-914'. cell_id: Target cell id. Use get_board_cells() to find cell ids.

Returns the updated card.

Inputs: card_ref (required), cell_id (required).

move_card_to_iterationA

Move a card into an iteration (sprint).

Args: card_ref: Card reference, e.g. 'ON-914'. iteration_id: Target iteration id. Use list_iterations() to find ids.

Returns the updated card.

Inputs: card_ref (required), iteration_id (required).

remove_card_labelA

Remove a single label from a card.

Args: card_ref: Card reference, e.g. 'ON-914'. label_id: Label id to remove.

Inputs: card_ref (required), label_id (required).

reopen_taskA

Reopen (un-complete) a task.

Args: card_ref: Card reference, e.g. 'ON-914'. task_id: Task id (from list_tasks).

Inputs: card_ref (required), task_id (required).

search_by_field_valueA

Find all cards where a custom field contains a given value.

Useful for finding cards tagged with a specific branch, PR URL, feature flag, or any other custom field content.

Args: field_id: Numeric custom field id from list_custom_fields(). value: Value to match (substring match, case-insensitive). limit: Maximum cards to scan (default 50).

Returns matching card objects.

Inputs: field_id (required), value (required), limit. Returns: result

search_cardsA

Full-text search across all cards in the organisation.

Searches card summaries, descriptions, and comments.

Args: query: Search string. Supports partial matches.

Returns a ranked list of matching cards with summary, reference, board name, and a short excerpt.

Inputs: query (required). Returns: result

set_card_fieldA

Set a single custom field on a card without touching other fields.

Args: card_ref: Card reference, e.g. 'ON-914'. field_id: Numeric custom field id. Use list_custom_fields() to look up ids. value: New field value (always a string).

Returns the updated card.

Inputs: card_ref (required), field_id (required), value (required).

unblock_cardA

Resolve (unblock) a specific blocker on a card.

If this was the last active blocker, the card's blocked flag is cleared. Use list_blockers() to find the blocker_id.

Args: card_ref: Card reference, e.g. 'ON-914'. blocker_id: Numeric blocker id from list_blockers(). resolution: Optional description of how the block was resolved.

Returns the resolved blocker object.

Inputs: card_ref (required), blocker_id (required), resolution.

update_cardA

Update one or more fields on an existing card.

Only the fields you provide will be changed — omitted fields are untouched.

Args: card_ref: Card reference, e.g. 'ON-914'. summary: New title. description: New description (markdown). points: New story points. due_date: Due date in YYYY-MM-DD format. Pass empty string to clear. iteration_id: Move card to this iteration/sprint id. Use list_iterations() to find ids. assignee_ids: Replace assignee list. Use list_members() to find ids. label_ids: Replace label list. extra_fields: Merge into existing custom fields. Existing keys not mentioned here will be preserved.

Returns the full updated card object.

Inputs: card_ref (required), summary, description, points, +5 more.

update_taskB

Update a task's description, completion state, or assignee.

Args: card_ref: Card reference, e.g. 'ON-914'. task_id: Task id (from list_tasks). description: New description (leave None to keep current). complete: True to complete, False to reopen. assignee_id: New assignee user id.

Inputs: card_ref (required), task_id (required), description, complete, +1 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 45 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose, covering cards, tasks, comments, labels, blockers, time tracking, activity logging, webhooks, and search. Only minor aliasing like find_card=get_card, which does not cause confusion.

    Naming Consistency4/5

    Most tools follow a consistent verb_noun pattern (create_card, list_tasks, etc.). However, there is some variation in prefix verbs (get vs list vs find vs search) and a few tools like log_activity and log_time use a different structure, preventing a perfect score.

    Tool Count3/5

    With 45 tools, the set is quite large but each tool serves a specific function within a comprehensive project management domain. It borders on excessive but remains justified given the breadth of features supported.

    Completeness3/5

    The tool surface covers most CRUD operations for core entities but misses lifecycle operations for iterations, epics, and boards (e.g., no create/delete). Card deletion is only via archiving, which is a notable gap.

    Maintenance

    ActivityNo data
    ResponsivenessSyncing

    Related MCP Connectors

    Related MCP Servers