| name | kickoff-task |
| description | Initialize or refresh a task note with interview-mode scoping. Questions ambiguities before planning. Defines execution and verification strategy. Activate when #kit appears anywhere in the user message. |
| argument-hint | [goal] |
| disable-model-invocation | false |
| allowed-tools | Read, Edit, MultiEdit, Write, Glob, Grep, Bash, Task |
Initialize task context for: $ARGUMENTS
Phase 1 — Discovery:
- Read repo-local
CLAUDE.md if it exists, plus the nearest build or test scripts.
- If
.claude/state/memory/index.md exists, read project knowledge for context on conventions and past decisions.
- Use
Explore or repo-librarian to quickly understand the relevant parts of the codebase.
- Check for existing test infrastructure (test frameworks, test commands, CI config).
Phase 2 — Scope interview (for ambiguous or complex goals):
- Before creating the plan, assess complexity:
- Trivial (single file, obvious fix) — skip interview, proceed directly.
- Simple (1-2 files, clear scope) — 1-2 targeted questions.
- Complex (3+ files, multiple interpretations) — full scoping interview.
- For complex goals, clarify:
- What should explicitly NOT be included? (scope boundaries)
- What constraints exist? (do not touch X, do not change Y)
- What does "done" look like? (acceptance criteria)
Phase 3 — Task note creation:
- Ensure
.claude/state/tasks/ exists in the current workspace.
- Create or refresh a task note under
.claude/state/tasks/ and point .claude/state/current-task.txt at it.
- Fill the note with:
- goal
- assumptions
- scope boundaries (what is included AND what is excluded)
- likely files or areas
- open questions
- test strategy (existing tests to run, new tests needed)
- planned verification
- next actions
- If the goal is vague, ask
planner-sisyphus for a short plan before filling the note.
- Reply with the task note path and the next concrete action.
