Stats
Actions
Tags
From superkit
Build one feature end-to-end with the combined spec-kit + superpowers loop — brainstorm, specify/clarify/plan/tasks, then worktree + TDD + subagent execution + review + finish.
How this skill is triggered — by the user, by Claude, or both
Slash command
/superkit:featureThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Drive one feature through the full combined loop. Run a dependency check first:
Drive one feature through the full combined loop. Run a dependency check first:
bash "${CLAUDE_PLUGIN_ROOT}/scripts/check-deps.sh" — if it fails, tell the user to run
/superkit:setup and stop.
Brainstorm — invoke the superpowers:brainstorming skill to refine intent and
produce/confirm a short design. (Skip only if the user already has a clear spec.)
Specify — invoke speckit-specify (spec-kit). It creates specs/<NNN>-name/spec.md
and a branch <NNN>-name. Capture the id and spec-path from its output.
Register — bash "${CLAUDE_PLUGIN_ROOT}/scripts/registry.sh" add <id> "<title>" - <spec-path> <branch>
then registry.sh set-status <id> specified.
Clarify — invoke speckit-clarify; then registry.sh set-status <id> clarified.
Plan — invoke speckit-plan; then registry.sh set-status <id> planned.
Tasks — invoke speckit-tasks; then registry.sh set-status <id> tasked.
Verification gate: confirm specs/<id>/tasks.md exists and is non-empty before executing.
Worktree — invoke superpowers:using-git-worktrees to get an isolated workspace
with a verified green baseline. registry.sh set-status <id> executing.
Execute — follow superpowers:test-driven-development (red-green-refactor) and
invoke superpowers:subagent-driven-development to implement specs/<id>/tasks.md
task-by-task. Do NOT re-plan — tasks.md is the contract. Pass each subagent ONLY its
task, not the whole tree. Then invoke superkit:iterate-to-green to drive the whole
suite to green (it asks for an iteration budget and handles stuck loops); do not proceed
to review until it reports green or you choose to stop.
Conformance + Review — invoke superkit:conformance-check for an advisory
acceptance-criteria checklist, then invoke superpowers:requesting-code-review.
registry.sh set-status <id> in-review.
Finish — invoke superpowers:finishing-a-development-branch (merge/PR + cleanup).
registry.sh set-status <id> done.
At each handoff, confirm the prior artifact exists / tests are green before continuing.
If a speckit-* skill is unavailable, fall back: Read .claude/skills/speckit-<step>/SKILL.md
and follow it, or ask the user to run /speckit-<step> and resume.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub joshuatrobertson/super-kit --plugin superkit