executing-plans

Executing Plans

Executing Plans is an agent skill from Superpowers in which your agent works through a written plan itself, task by task with tests, then gets one fresh review of the whole branch.

Checked by CookiesForAI on : license and last change of the skill re-read from GitHub.

Written by

Install Executing Plans

Install only this skill with the skills CLI

Terminal
npx skills add obra/superpowers --skill executing-plans

Or, in Claude Code, install the whole Superpowers plugin from Anthropic's official marketplace

Agent chat
/plugin install superpowers@claude-plugins-official

Or paste this into your coding agent: Install the agent skill executing-plans from github.com/obra/superpowers. Checked by CookiesForAI on : license and last change of the skill re-read from GitHub. A skill can include scripts that run on your computer, so read its source first.

What Executing Plans does

Executing Plans is one of the skills in Superpowers, Jesse Vincent's open-source method for AI coding agents. It loads when the agent carries out an implementation plan itself in the current session, because you chose inline execution or because no subagent tool is available. One context does every task, and one fresh reviewer checks the whole branch at the end.

Under Executing Plans, work happens in an isolated workspace from Using Git Worktrees, and Test-Driven Development governs every step: each test is watched failing before the code that makes it pass. When the code is wrong, the agent switches to Systematic Debugging. Progress goes into a ledger file that survives context compaction and is shared with Subagent-Driven Development. Bundled task-start and task-done scripts print each task's brief and mark a task complete only when its tests pass.

Under Executing Plans, the agent does not stop to check in between tasks. It decides plan conflicts itself and records each as a ruling, stopping only for destructive or security-sensitive actions, a merge, push or publish, or a plan too broken to follow. The final review uses the reviewer template from Requesting Code Review. The agent then lists every ruling and deferred minor finding for you and hands off to Finishing a Development Branch.

When to use Executing Plans

  • You have a finished implementation plan and want the agent to carry it out without asking after every task.
  • Your agent tool has no subagents, or you want the agent to do every task itself.
  • You want a written record of every decision the agent made on your behalf while following the plan.

When to pick something else

  • Your plan has many independent tasks and you want each one reviewed before the next: use Subagent-Driven Development.

Which agents Executing Plans works in

Jesse Vincent documents Executing Plans for Claude Code, Codex (app and CLI), Cursor, Gemini CLI, GitHub Copilot CLI, OpenCode, Antigravity, Devin CLI, Factory Droid, Grok Build CLI, Kimi Code, Pi, Qwen Code, Hermes Agent, Muse. The open skills CLI also installs it into 78 agents, including Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode (we listed it with the CLI on ). See where each agent looks for skills.

Superpowers is written in Claude Code's vocabulary (subagent dispatch, todos, the Skill tool) and ships tool mappings for Codex, Gemini CLI, Pi, Antigravity, Hermes Agent and Muse. Installed as the full plugin, a session-start hook loads Using Superpowers so the agent reaches for skills on its own; one skill installed with the skills CLI gets no hook and may point to sibling skills you do not have.

Executing Plans license

Executing Plans is published under MIT.

More from Jesse Vincent

Lists that include executing-plans