PWR Sprint
A board that reflects real state
PWR Sprint is a Kanban board and sprint system built on the task state machine, with connectors for Jira and GitHub Issues. Sprint Autopilot analyzes a sprint, proposes what agents can take, and executes only the tickets you approve.
Team leads and producers who run sprints and want agents to take approved tickets without losing track of status.
How it works
PWR Sprint, step by step
- 01
Bring in the work
Create tasks directly, or import issues from Jira or GitHub Issues. Imported issues always arrive as drafts.
- 02
Plan the sprint
Add tasks to a sprint with a credit budget and a parallel-run limit. Sprint Autopilot can analyze an external sprint and propose which tickets are ready for automation.
- 03
Approve what agents may take
Eligibility is explicit: an approval label on the issue or an approval in PWR. Tickets without acceptance criteria are refused.
- 04
Follow progress
Cards move as tasks and pull requests change state. The sprint report is built from recorded transitions and metered credits.
- Available
- Implemented and covered by automated tests in the current platform build.
- Beta
- Implemented; verification against the live external service is still pending.
- Upcoming
- Planned or in progress. Not usable yet.
Kanban board
The board is a view over task status and workflow stage. It never stores a status of its own.
State-driven columns
AvailableCards are placed by task status and workflow stage. A running task that is being validated or checked by CI sits in Testing because the worker put it there.
Drag-and-drop through real transitions
AvailableMoving a card performs the underlying task transition with your role's permissions. Moves with no legal path are refused with the allowed targets.
Configurable columns
AvailableAdmins define columns per project. Every task status must map to exactly one column, so a layout cannot hide tasks. WIP limits are flagged, not enforced.
Cards with evidence
AvailableEach card shows its agent, dependencies, blockers, branch, pull request, validation and CI status, and estimated and actual credits. Unknown values are empty, not guessed.
Branch, pull request and CI fields come from the GitHub workflow, which is in beta.
Sprints
Sprint lifecycle
AvailablePlanned, active, completed and cancelled sprints, with one active sprint per project and no double-booking of a task across open sprints.
Reports from real data
AvailableBurndown by task count from recorded completions, tasks by status and column, blockers, pull requests and credits consumed. Nothing is projected or interpolated.
Sprint budgets
AvailableA credit budget and a maximum number of concurrent runs per sprint. Execution refuses to start without a budget.
Sprint execution
BetaSelect tasks in an active sprint and execute them as a goal. Tasks are queued in dependency order within the budget, and each one goes through validation, review and human approval.
Controller logic is covered by automated tests with a stub agent and a simulated GitHub; live end-to-end verification is pending.
Project management connectors
Connectors import work and write back recorded facts, not model output.
Jira Cloud
BetaImport by sprint, board, JQL or all open issues. Maps priority, links, labels and acceptance criteria. Authenticates with an API token.
Verified against a simulated Jira Cloud; verification against a live Jira site is pending.
GitHub Issues
BetaImport open issues by milestone or label from the connected repository. Milestones are treated as sprints.
Verified against a simulated GitHub; verification against live GitHub Issues is pending.
Verified write-back
BetaComments, transitions, labels and pull request links are written back from recorded task status, validation result and review verdict. Each write is idempotent and audited.
Verified against simulated Jira and GitHub; live verification is pending.
Conflict detection
BetaIf a task was edited in PWR and the issue changed externally, the conflict is recorded and the ticket is held back from automation until a person resolves it.
Verified against simulated Jira and GitHub; live verification is pending.
Sprint Autopilot
BetaImports an external sprint, has the Producer role assess each ticket, and presents a proposal. Only the tickets you approve are executed.
Verified against simulated trackers with fabricated model replies; live verification is pending.
Linear and Notion
UpcomingConnector interfaces are declared but not implemented.
Questions
Will agents pick up any ticket in my backlog?
No. Imported issues become draft tasks. A ticket is eligible for autonomous execution only after an explicit signal: the approval label on the issue or an approval in PWR. It also needs acceptance criteria.
What does PWR write back to Jira?
Recorded facts only: the task status, the pull request link, whether validation passed, and the review verdict. Transitions and labels are written only on two-way links with a status map you define. PWR does not create Jira issues.
Does the board show progress percentages or ETAs?
No. The board and the sprint report show recorded state: status, stage, completions per day and credits consumed. There are no derived percentages or forecasts.
Does Sprint Autopilot merge the work it executes?
No. Each executed ticket produces a pull request that goes through validation, independent review and a human approval before the merge gate.
Has the Jira connector been used with a real Jira site?
Not yet. It is implemented against the Jira Cloud REST and Agile APIs and tested against a simulator. It is marked beta until it has been verified on a live site.
Put an AI development team on your idea.
Describe what you want to build. PWR plans, builds, tests and reviews it, and you approve each step that matters.