A community-maintained fork of obra/superpowers specifically for Claude Code users.
The original Superpowers is designed as a cross-platform toolkit that works across multiple AI CLI tools (Claude Code, Codex, OpenCode, Gemini CLI). Features unique to Claude Code fall outside the scope of the upstream project due to its cross-platform nature.
This fork integrates Claude Code-native features into the Superpowers workflow.
- Leverage Claude Code-native features as they're released
- Community-driven - contributions welcome for any CC-specific enhancement
- Track upstream - stay compatible with obra/superpowers core workflow
| Feature | Claude Code Version | Description |
|---|---|---|
| Native Task Management | v2.1.16+ | Dependency tracking, real-time progress visibility |
| Structured Task Metadata | v2.1.16+ | Goal/Files/AC/Verify structure with embedded json:metadata |
| Pre-commit Task Gate | v2.1.16+ | Plugin hook blocks git commit when tasks are incomplete |
| Superpowers (Vanilla) | Superpowers Extended CC |
|---|---|
|
|
|
# Register marketplace
/plugin marketplace add pcvelz/superpowers
# Install plugin
/plugin install superpowers-extended-cc@superpowers-extended-cc-marketplace/plugin install --source url https://github.com/pcvelz/superpowers.gitCheck that commands appear:
/help# Should see:
# /superpowers-extended-cc:brainstorming - Interactive design refinement
# /superpowers-extended-cc:writing-plans - Create implementation plan
# /superpowers-extended-cc:executing-plans - Execute plan in batches
-
brainstorming - Activates before writing code. Refines rough ideas through questions, explores alternatives, presents design in sections for validation. Saves design document.
-
using-git-worktrees - Activates after design approval. Creates isolated workspace on new branch, runs project setup, verifies clean test baseline.
-
writing-plans - Activates with approved design. Breaks work into bite-sized tasks (2-5 minutes each). Every task has exact file paths, complete code, verification steps. Creates native tasks with dependencies.
-
subagent-driven-development or executing-plans - Activates with plan. Dispatches fresh subagent per task with two-stage review (spec compliance, then code quality), or executes in batches with human checkpoints.
-
test-driven-development - Activates during implementation. Enforces RED-GREEN-REFACTOR: write failing test, watch it fail, write minimal code, watch it pass, commit. Deletes code written before tests.
-
requesting-code-review - Activates between tasks. Reviews against plan, reports issues by severity. Critical issues block progress.
-
finishing-a-development-branch - Activates when tasks complete. Verifies tests, presents options (merge/PR/keep/discard), cleans up worktree.
The agent checks for relevant skills before any task. Mandatory workflows, not suggestions.
When writing-plans creates tasks, each task carries structured metadata that survives across sessions and subagent dispatch:
TaskCreate:
subject: "Task 1: Add price validation to optimizer"
description: |
**Goal:** Validate input prices before optimization runs.
**Files:**
- Modify: `src/optimizer.py:45-60`
- Create: `tests/test_price_validation.py`
**Acceptance Criteria:**
- [ ] Negative prices raise ValueError
- [ ] Empty price list raises ValueError
- [ ] Valid prices pass through unchanged
**Verify:** `pytest tests/test_price_validation.py -v`
```json:metadata
{"files": ["src/optimizer.py", "tests/test_price_validation.py"],
"verifyCommand": "pytest tests/test_price_validation.py -v",
"acceptanceCriteria": ["Negative prices raise ValueError",
"Empty price list raises ValueError",
"Valid prices pass through unchanged"]}
```The json:metadata block is embedded in the description because TaskGet returns the description but not the metadata parameter. This ensures metadata is always available — for executing-plans verification, subagent-driven-development dispatch, and .tasks.json cross-session resume.
Testing
- test-driven-development - RED-GREEN-REFACTOR cycle (includes testing anti-patterns reference)
Debugging
- systematic-debugging - 4-phase root cause process (includes root-cause-tracing, defense-in-depth, condition-based-waiting techniques)
- verification-before-completion - Ensure it's actually fixed
Collaboration
- brainstorming - Socratic design refinement + native task creation
- writing-plans - Detailed implementation plans + native task dependencies
- executing-plans - Batch execution with checkpoints
- dispatching-parallel-agents - Concurrent subagent workflows
- requesting-code-review - Pre-review checklist
- receiving-code-review - Responding to feedback
- using-git-worktrees - Parallel development branches
- finishing-a-development-branch - Merge/PR decision workflow
- subagent-driven-development - Fast iteration with two-stage review (spec compliance, then code quality)
Meta
- writing-skills - Create new skills following best practices (includes testing methodology)
- using-superpowers - Introduction to the skills system
- Test-Driven Development - Write tests first, always
- Systematic over ad-hoc - Process over guessing
- Complexity reduction - Simplicity as primary goal
- Evidence over claims - Verify before declaring success
Read more: Superpowers for Claude Code
Contributions for Claude Code-specific enhancements are welcome!
- Fork this repository
- Create a branch for your enhancement
- Follow the
writing-skillsskill for creating and testing new skills - Submit a PR
See skills/writing-skills/SKILL.md for the complete guide.
Claude Code may automatically enter Plan mode during planning tasks, which conflicts with the structured skill workflows in this plugin. To prevent this, add EnterPlanMode to your permission deny list.
In your project's .claude/settings.json:
{
"permissions": {
"deny": ["EnterPlanMode"]
}
}This blocks the model from calling EnterPlanMode, ensuring the brainstorming and writing-plans skills operate correctly in normal mode. See upstream discussion for context.
When using native tasks, the agent should not commit until all tasks are finished. This plugin includes an example hook that blocks git commit when tasks are still open.
Add this to your .claude/settings.local.json:
{
"hooks": {
"PreToolUse": [
{
"matcher": "Bash",
"hooks": [
{
"type": "command",
"command": "bash ~/.claude/plugins/marketplaces/superpowers-extended-cc-marketplace/hooks/examples/pre-commit-check-tasks.sh"
}
]
}
]
}
}The hook ships with the plugin at hooks/examples/pre-commit-check-tasks.sh. The marketplace path is stable across versions. It parses the session transcript for TaskCreate/TaskUpdate calls and blocks git commit when any tasks are not completed, cancelled, or deleted. Non-commit Bash commands pass through unaffected.
Skills update automatically when you update the plugin:
/plugin update superpowers-extended-cc@superpowers-extended-cc-marketplaceThis fork tracks obra/superpowers main branch. Changes specific to Claude Code are additive - the core workflow remains compatible.
MIT License - see LICENSE file for details