| name | sprint |
| description | Human-supervised development cycle — orchestrates roadmap → claim → autoplan → implement → commit for one skill issue |
| argument-hint | [issue-number] |
| allowed-tools | Bash, Read, Skill |
| automation | manual |
| user-invocable | true |
| metadata | {"version":"1.0","created":"2026-04-28T00:00:00.000Z","author":"Ability.ai","changelog":["1.0: Initial version — human-supervised cycle orchestrating roadmap → claim → autoplan → implement → commit for one skill issue"]} |
Sprint
ℹ️ First, set expectations: before anything else, print one short line with this skill's version and its most recent change — the top entry of metadata.changelog above — e.g. sprint vX.Y — recent: <summary>. Then proceed.
Walks through one complete agent development cycle for a single issue. The implement step is intentionally human-driven — this skill frames and bookends it. The autonomous equivalent is /work-loop.
When to Use
- You want to work on a specific backlog issue with full context at each step
- You want autoplan analysis before touching any skill files
- You're onboarding to the agent and want a guided workflow
Process
Step 1: Roadmap Check
Invoke /roadmap to show the skill-grouped backlog. This surfaces which skills have the most open work and helps choose a focus area.
If $ARGUMENTS is provided (a specific issue number), skip roadmap and go directly to Step 2 with that issue.
Step 2: Claim
Invoke /claim (or /claim $ARGUMENTS if an issue number was provided).
If already in-progress from a previous session, continue with that issue and skip to Step 3.
Step 3: Autoplan
Invoke /autoplan on the claimed issue.
The plan will identify:
- Which SKILL.md is affected
- Which section to change
- The right tool to use (adjust-playbook vs create-playbook)
- Any risks
Wait for the plan output before proceeding.
Step 4: Implement (Human Step)
Print clearly:
## Ready to Implement
Based on the autoplan, run one of:
/adjust-playbook $SKILL_NAME — to modify an existing skill
/create-playbook — to scaffold a new skill
Make your changes, then return here and confirm to commit.
Ask the user: "Type 'done' when the implementation is complete, or 'abort' to stop without committing."
If 'abort': do nothing. Leave the issue in-progress for the next session.
Step 5: Commit
Invoke /commit to stage the changed skill files, write the commit message, and close the issue.
Step 6: Finish
## Sprint Complete
Issue #$NUMBER closed.
Run `/sprint` again to start the next issue, or `/backlog` to review remaining work.
Outputs
- One issue moved from todo → done
- One git commit with issue traceability
- Full context at each step before acting
Notes
- Sprint always works one issue at a time — for bulk processing, use
/work-loop
- If you want to skip straight to committing (already implemented), run
/commit directly
- If you want to plan without claiming, run
/autoplan $ISSUE_NUMBER directly