From sprint
Add backlog items to an existing sprint through guided brainstorming and planning. Automatically detects sprint context or asks for sprint location. Includes brainstorming phase, plan mode for task design, and ensures tasks are sized for single Claude Code sessions. Triggers include "add backlog", "new feature", "add task", "백로그 추가", "기능 추가", "태스크 추가".
How this skill is triggered — by the user, by Claude, or both
Slash command
/sprint:add-backlogThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Add work items to a sprint through guided brainstorming and structured planning.
Add work items to a sprint through guided brainstorming and structured planning.
This skill guides you through:
Key constraint: Each Task must be completable in a single Claude Code session.
.sprint-config in sprint directory"mode": "worktree":
"mode": "default" or no config:
Use conversational discovery to understand what to build.
Principles:
Discovery flow:
What are we building?
Type identification
coding - Software development, writing codedocs - Documentation, guides, technical writingideation - Brainstorming, exploring ideasgeneral - Research, planning, other tasksScope clarification
Success criteria
Load type-specific guidelines
assets/{type}/guidelines.mdEnter Plan Mode to design the work structure.
Feature slug assignment:
user-auth)F-{feature-slug} (e.g., F-user-auth)Task slug assignment:
login-api)T-{task-slug} (e.g., T-login-api); a Sub-task is T-{task-slug}.{sub-slug} (e.g., T-login-api.schema)No sequential numbering. Features and Tasks are identified by descriptive slugs, not
max+1numbers.
Task sizing rule (CRITICAL):
Each Task must be completable in a single Claude Code session. If too big, break into smaller Tasks.
Ask yourself: "Can an agent complete this without session reset?"
- Yes → Good size
- No → Break it down further
Design the structure:
### F-{feature-slug}: {Feature Name}
> {Brief description}
**Design**: [refs/designs/F-{feature-slug}.md](refs/designs/F-{feature-slug}.md)
**Context**: [active/F-{feature-slug}.md](active/F-{feature-slug}.md)
- [ ] T-{task-slug}: {Task name} `backlog`
- [ ] T-{task-slug}.{sub-slug}: {Sub-task if needed}
- [ ] T-{another-slug}: {Task name} `backlog`
- [ ] T-review-{feature-slug}: Review & Refactor {Feature Name} `backlog` ← verification/refactoring
Note: Type, Goals, and detailed design information go in the Design doc, not BACKLOG.md.
Review & Refactor tasks (REQUIRED):
Task cycle: ideation → plan → implement → verify/refactor → ... → done → final review/refactor
| Situation | Task to add |
|---|---|
| End of every Feature | Review & Refactor {Feature Name} (slug T-review-{feature-slug}) - verify, improve, clean up |
| After complex Task (3+ sub-tasks) | Review {task name} (slug T-review-{task-slug}) - intermediate check & improve |
| Integration points | Review integration (slug T-review-integration) - verify + refactor if needed |
R&R Tasks must keep "Review & Refactor" in the title —
/sprint:work-on-featureidentifies the final R&R Task by its title, not its slug.
What Review & Refactor includes:
Examples:
- [ ] T-login-api: Login API `backlog`
- [ ] T-signup-api: Signup API `backlog`
- [ ] T-review-auth-apis: Review login/signup APIs `backlog` ← intermediate verification
- [ ] T-token-refresh: Token refresh `backlog`
- [ ] T-review-user-auth: Review & Refactor User Auth `backlog` ← end-of-Feature review/refactor
Present to user:
When sprint is in worktree mode:
Read .sprint-config → get worktree.command and worktree.pathPrefix
Worktree for this Feature:
git worktree listIf creating new worktree:
.sprint-config (worktree.command)
feature/F-{feature-slug}worktree.pathPrefix + F-{feature-slug}Record in Feature:
Branch: and Worktree: fields to Feature in BACKLOG.mdEvery Feature in worktree mode MUST have a final merge Task:
- [ ] T-merge-{feature-slug}: Merge {branch-name} and clean up worktree `backlog`
This Task:
git worktree remove)Task order:
- [ ] T-login-api: ... `backlog`
- [ ] T-signup-api: ... `backlog`
- [ ] T-review-user-auth: Review & Refactor User Auth `backlog` ← review/refactor
- [ ] T-merge-user-auth: Merge feature/F-user-auth and clean up worktree `backlog` ← merge (last)
When designing the Feature structure, evaluate whether Tasks would benefit from agent teams:
For speed:
For quality:
If applicable, include in the proposal to user:
If agent teams are not applicable (sequential dependencies or simple scope):
Agent teams execute during
@INSTRUCTION.md, not during this skill. This step only identifies opportunities and documents them in the design.
After user approval:
Update BACKLOG.md
Create Design doc (refs/designs/F-{feature-slug}.md)
Create Active context (active/F-{feature-slug}.md)
Create worktree (worktree mode, if "create new" selected)
.sprint-config)Confirm addition
Added to BACKLOG.md:
- F-{feature-slug}: {Feature Name}
- T-{task-slug}: {Task}
- T-{another-slug}: {Task}
Created:
- refs/designs/F-{feature-slug}.md (design doc)
- active/F-{feature-slug}.md (working context)
Ready to work? Start with: @INSTRUCTION.md #your-name
After confirming addition, offer to detail-plan Tasks:
Would you like to detail-plan any of the new Tasks now?
1. Yes → Which Task? (runs /sprint:plan-backlog)
2. No → End skill
If user selects Yes:
/sprint:plan-backlog for the selected TaskIf user selects No:
/sprint:plan-backlog later when neededType-specific guidelines:
assets/coding/guidelines.md - For software development tasksassets/docs/guidelines.md - For documentation tasksassets/ideation/guidelines.md - For brainstorming/exploration tasksassets/general/guidelines.md - For general tasksCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub workingdanny911/sprint --plugin sprint