Start with any shape of work
Use a simple bucket of tagged tasks, a list, cards, or custom swimlanes. The system should adapt to the team instead of forcing a ceremony.
Quick start and knowledgebase
Quick start: build a planning system that works for humans first and agents when the task is ready.
The fastest path is simple: capture work in whatever shape the team understands, then progressively add contracts, context, policies, time tracking, reminders, and evidence requirements until the work is safe for humans or agents to execute.
How to start
From loose task bucket to governed agent execution.
Customize statuses and swimlanes
Admins can model backlog, ready, in progress, review, done, customer validation, human blocked, or any other lane the workspace needs.
Define the task contract
Add scope, non-goals, constraints, expected output, evidence requirements, approval rules, and allowed tools before a human or agent starts.
Attach context packs
Give agents the exact files, docs, repo hints, decisions, and environment notes needed to complete the task without guessing.
Turn reminders into quality control
Warn when tickets have no expected completion date, no recent update, overdue due dates, missing evidence, or time logged over estimate.
Let agents connect through MCP
Codex, Claude Code, Cursor, Cowork, and other compatible tools can discover ready work and fetch task contracts through the remote MCP endpoint.
Migration playbook
Keep the human workflow. Add the agent execution layer.
Teams moving from Jira, Linear, or Azure DevOps should not throw away working habits. Preserve the useful tracking data, then add task contracts, context packs, execution policy, readiness queues, and Evidence-based Done before agents start work.
- Sources
- 3
- Setup steps
- 12
- Agent upgrades
- 9
From Jira
Deep workflow configuration, boards, backlog grooming, due dates, automation rules, comments, links, and work logs.
Keep
- Issue keys, summaries, descriptions, priority, assignee, labels, links, due dates, estimates, and comments.
- Board columns and workflow states that people already understand.
- Automation intent such as assignment, linking, work logging, and escalation triggers.
buildr-plannr upgrade
- Convert vague Jira descriptions into task contracts with scope, non-goals, allowed tools, acceptance checks, and rollback notes.
- Turn linked Confluence, repo, support, and architecture references into context packs.
- Replace status-only completion with Evidence-based Done: tests, logs, screenshots, PRs, risk notes, and reviewer acceptance.
Setup path
- Import active work first, then archive or export stale history separately.
- Map Jira statuses to visible swimlanes and keep custom blocker lanes such as Needs human input.
- Create saved views for human triage, agent-ready work, missing context, and evidence review.
- Require task contracts and context packs before any imported issue can enter Ready for agent.
Agent readiness gate
Imported Jira work is agent-ready only after the contract, context pack, execution policy, and evidence requirements are complete.
Execution control
Use scoped MCP/API tokens and approval gates before agents mutate imported Jira work.
From Linear
Fast issue capture, team workflows, display options, imports, triage, relations, projects, cycles, and lightweight views.
Keep
- Issue titles, descriptions, workflow states, priority, estimates, labels, comments, relations, projects, initiatives, and dashboards where available.
- Team-specific workflow language such as Backlog, Todo, In Progress, In Review, Done, and Canceled.
- Clean display options such as grouping by status, assignee, project, priority, cycle, or focus.
buildr-plannr upgrade
- Add agent-specific readiness queues instead of relying only on human workflow status.
- Persist workspace planning defaults so humans and MCP clients see the same custom statuses, fields, and views.
- Attach execution policies, approval history, reminder policy, and time signals to each issue before agent execution.
Setup path
- Run a small pilot import and verify which concepts should be kept versus left behind.
- Map Linear workflow states into buildr-plannr swimlanes and saved views.
- Create one human planning view and one agent execution view before inviting the wider team.
- Require evidence and approvals before imported issues can move from In Review to Done.
Agent readiness gate
Imported Linear work is agent-ready only when the issue explains what an agent may do, what context is trusted, and what proof is required.
Execution control
Use readiness queues, run quota, and MCP workspace customization so coding agents do not assume the old Linear workflow shape.
From Azure DevOps
Structured boards, work items, backlogs, sprints, queries, delivery plans, analytics, and Microsoft ecosystem integration.
Keep
- Epics, features, user stories, tasks, bugs, owners, priorities, iteration intent, links, and delivery-plan dependencies.
- Queries that teams use for triage, bulk updates, dashboards, and cross-team reporting.
- Sprint and backlog planning habits that give managers delivery visibility.
buildr-plannr upgrade
- Keep portfolio structure, then add agent execution contracts at the issue level.
- Turn query-driven triage into readiness queues for missing context, blocked work, approvals, and evidence review.
- Use proof requirements and MCP permissions to make agent execution auditable outside the Microsoft suite.
Setup path
- Choose the backlog levels that should become projects, milestones, or issue links.
- Map Azure work item states into human swimlanes and agent readiness lanes.
- Convert delivery-plan risks into dependencies, blockers, target dates, and context pack requirements.
- Rebuild critical queries as saved views for human triage, agent claims, overdue work, and evidence gaps.
Agent readiness gate
Imported Azure Boards work is agent-ready only after dependencies, target dates, context, policy, and verification evidence are explicit.
Execution control
Use execution policy and approval gates when agents touch repositories, environments, billing surfaces, or customer-facing work.
MCP quickstart
Remote MCP endpoint
Give coding tools a governed way to ask what to do next. The remote MCP surface starts at /api/mcp. Agents authenticate with scoped planner API tokens and fetch workspace customization, ready work, task contracts, context packs, issue search, status updates, comments, time logs, evidence, and approvals without leaving their coding tool.
initialize
MCP JSON-RPC method supported by the planner endpoint.
tools/list
MCP JSON-RPC method supported by the planner endpoint.
tools/call
MCP JSON-RPC method supported by the planner endpoint.
Client setup guides
Connect Codex, Claude Code, Cursor, Cowork, or any MCP client.
- Clients
- 5
- Capabilities
- 21
- Write-ready
- 4
Codex
Coding agents working in a repo checkout.
Use an agent-scoped token for autonomous runs and a user-scoped token for assisted human sessions.
Setup
- Create a scoped agent API token in workspace settings.
- Register the remote server URL as https://plannr.buildrlab.com/api/mcp or the dev URL for non-production testing.
- Pass the token with either Authorization: Bearer <token> or x-api-key.
Verify
- Call initialize and confirm the buildr-plannr server info.
- Call tools/list and confirm planner tools are visible.
- Call tools/call with buildr_plannr.get_workspace_customization before moving issues so custom swimlanes, fields, WIP limits, and evidence policy are respected.
Capabilities: workspace customization, ready work discovery, agent claim queue, claim leases, agent status posts, claim cancellation, task contracts, context packs, comments, time logging, evidence.
Claude Code
Terminal-based coding assistants that can call remote MCP tools.
Prefer a narrow agent token with read-work, read-context-pack, comment, and evidence scopes first.
Setup
- Add the buildr-plannr remote MCP URL to the client configuration.
- Store the planner token in the client secret store or environment.
- Keep write-capable scopes off until the workspace policy allows them.
Verify
- Call initialize and confirm the buildr-plannr server info.
- Call tools/list and inspect the input schema before the first run.
- Call tools/call with buildr_plannr.get_workspace_customization and cache the current status, field, reminder, and evidence rules.
Capabilities: workspace customization, tool discovery, claim leases, task contracts, agent comments, approval requests, evidence.
Cursor
IDE users who want issue context next to code changes.
Use per-workspace tokens; rotate them when an IDE workspace is shared or handed over.
Setup
- Add the remote MCP endpoint to the workspace-level MCP configuration.
- Use a token scoped to the current workspace and agent profile.
- Keep the planner issue ID in the task prompt so comments and evidence attach to the right work item.
Verify
- Call initialize and confirm the buildr-plannr server info.
- Call tools/list and confirm planner tools are visible.
- Call tools/call with buildr_plannr.search_issues from the IDE.
Capabilities: workspace customization, issue search, claim leases, agent status posts, context packs, status updates, time logging, approval requests.
Cowork
Team agents that claim and execute queued work.
Use an agent token pinned to the Cowork agent ID so it cannot post as another actor.
Setup
- Create a dedicated agent identity for Cowork.
- Grant only the issue scopes needed by that agent role.
- Configure the remote endpoint and token in the agent runtime.
Verify
- Call initialize and confirm the buildr-plannr server info.
- Call tools/list and confirm planner tools are visible.
- Call tools/call with buildr_plannr.get_workspace_customization and confirm the active swimlane names match the planning board.
Capabilities: workspace customization, ready work discovery, agent claim queue, claim leases, claim cancellation, status updates, comments, approval requests, evidence.
Generic MCP client
Any MCP-compatible client that can call streamable HTTP JSON-RPC.
Choose the narrowest token scope that allows the client to perform its exact workflow.
Setup
- Set the server URL to /api/mcp.
- Send JSON-RPC 2.0 requests over HTTPS.
- Pass a scoped token and keep client logs redacted.
Verify
- Call initialize.
- Call tools/list.
- Call tools/call with buildr_plannr.get_workspace_customization before any write-capable workflow.
Capabilities: workspace customization, initialize, tools/list, tools/call, claim leases, search, read-only onboarding.
Knowledgebase
Ask how to set up human and agent workflows.
This finder uses curated static answers and keyword matching only. It does not send questions to an AI model, so teams can rely on repeatable onboarding guidance.
Guide
How is buildr-plannr different from Jira, Azure DevOps, or Linear?
Jira, Azure DevOps, and Linear track work for humans. buildr-plannr prepares work so AI agents can safely execute it: every issue can carry a task contract, context pack, execution policy, approval history, evidence requirements, time signals, and remote MCP access from the same source of truth.
- Start with normal issues, tags, lists, cards, and swimlanes.
- Add the task contract, context pack, and execution policy when work should become executable.
- Require tests, logs, screenshots, PRs, risk notes, and approvals before review or done.
Guide
How do I customize statuses, swimlanes, and work buckets?
Use workspace statuses to model your real process: backlog, ready, ready for agent, in progress, needs human input, review, customer validation, done, or any custom lane your team needs. Saved views and filters let each user keep a board, list, or tag bucket that matches their job.
- Open the planner workbench and choose the lanes that should be visible.
- Create or select a saved view for the workflow you want.
- Switch between board, list, and tag views depending on the job.
Guide
What should I configure first after creating a workspace?
Start with the smallest useful workflow: choose the visible swimlanes, create one saved view for human work, create one saved view for agent-ready work, define the reminder policy, then add API tokens only for tools that need MCP access.
- Open workspace settings and confirm support, roles, imports, notifications, audit, and MCP clients.
- Open the issue workbench and choose the swimlanes that match the team's current process.
- Save a human planning view and an agent execution view before inviting more users.
Guide
How do we migrate from Jira, Linear, or Azure DevOps?
Migrate in two passes: keep the useful human tracking data first, then add the agent execution layer. Import active issues, map familiar statuses into swimlanes and saved views, preserve useful links and comments, then require task contracts, context packs, execution policy, and evidence gates before imported work becomes agent-ready.
- Import active work first and keep stale history as archive or export evidence.
- Map source statuses, boards, queries, and display options into buildr-plannr swimlanes and saved views.
- Create human triage, agent-ready, missing-context, and evidence-review views before inviting the wider team.
- Require task contracts, context packs, approval policy, and Evidence-based Done before agents claim imported work.