Community · by obra

Executing plans (Superpowers) Claude skill

The Executing plans (Superpowers) skill helps an AI assistant carry out an implementation plan task by task within the current session, working as the implementer itself.

MIT licence
★ 301,993 on GitHub (whole repository)
Verified against its source 19 Oct 2026

View SKILL.md Repository

What the Executing plans (Superpowers) skill does

It lets the assistant execute a plan's tasks itself in the current session rather than handing each task to a separate implementer or reviewer subagent. The quotes say each step is proven with a test that is watched failing and then passing, and that decisions are recorded in a ledger. The skill is used when running an implementation plan in the current session with the assistant as the implementer.

What it covers

Inline task-by-task plan execution in the current session
Test-driven development per task step
Recording rulings on plan conflicts and deviations
One final whole-branch review and a single fix pass for Critical and Important findings

How to install it

In Claude Code

/plugin install superpowers@claude-plugins-official

This plugin also installs: brainstorming, diagnosing-superpowers, dispatching-parallel-agents, finishing-a-development-branch, receiving-code-review, requesting-code-review, subagent-driven-development, systematic-debugging, test-driven-development, using-git-worktrees, using-superpowers, verification-before-completion, writing-plans, writing-skills.

Steps change between versions. Check the official instructions before you install.