|
|
vor 9 Monaten | |
|---|---|---|
| .claude | vor 9 Monaten | |
| .claude-plugin | vor 10 Monaten | |
| .codex | vor 9 Monaten | |
| .github | vor 9 Monaten | |
| .opencode | vor 9 Monaten | |
| agents | vor 9 Monaten | |
| commands | vor 10 Monaten | |
| docs | vor 9 Monaten | |
| hooks | vor 10 Monaten | |
| lib | vor 9 Monaten | |
| skills | vor 9 Monaten | |
| tests | vor 9 Monaten | |
| .gitignore | vor 11 Monaten | |
| LICENSE | vor 11 Monaten | |
| README.md | vor 9 Monaten | |
| RELEASE-NOTES.md | vor 9 Monaten |
AI agents skip steps under time pressure. They bypass best practices when confident. They lack consistency across tasks. The result: bugs you didn't catch, designs you didn't validate, tests you didn't write.
Superpowers fixes this. Skills are mandatory instruction documents agents must follow. When a relevant skill exists, the agent checks for it, uses it, or fails the task.
At session start, the agent learns which skills exist. Before any task, the agent checks: "Does a skill match this work?" If yes, the agent loads and follows that skill.
Skills are markdown files with proven workflows. The test-driven-development skill forces RED-GREEN-REFACTOR. No test-first? Delete the code and start over. The skill prevents rationalization.
When you ask to build a feature:
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.
subagent-driven-development or executing-plans - Activates with plan. Dispatches fresh subagent per task (same session, fast iteration) or executes in batches (parallel session, 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.
Note: Installation differs by platform. Claude Code has a built-in plugin system. Codex and OpenCode require manual setup.
In Claude Code, register the marketplace first:
/plugin marketplace add obra/superpowers-marketplace
Then install the plugin from this marketplace:
/plugin install superpowers@superpowers-marketplace
Check that commands appear:
/help
# Should see:
# /superpowers:brainstorm - Interactive design refinement
# /superpowers:write-plan - Create implementation plan
# /superpowers:execute-plan - Execute plan in batches
Tell Codex:
Fetch and follow instructions from https://raw.githubusercontent.com/obra/superpowers/refs/heads/main/.codex/INSTALL.md
Detailed docs: docs/README.codex.md
Tell OpenCode:
Fetch and follow instructions from https://raw.githubusercontent.com/obra/superpowers/refs/heads/main/.opencode/INSTALL.md
Detailed docs: docs/README.opencode.md
Testing (skills/testing/)
Debugging (skills/debugging/)
Collaboration (skills/collaboration/)
Meta (skills/meta/)
Read more: Superpowers for Claude Code
Skills live directly in this repository. To contribute:
writing-skills skill for creating new skillstesting-skills-with-subagents skill to validate qualitySee skills/meta/writing-skills/SKILL.md for the complete guide.
Skills update automatically when you update the plugin:
/plugin update superpowers
MIT License - see LICENSE file for details