Skip to content
Get packs

Executing Plans Inline

Lets the agent carry out a written implementation plan itself, task by task, with a ledger and tests.

Source Download skill
Repository
jnMetaCode/superpowers-zh
Stars
8,444
Licence
MIT
Files
1 file

What it does

When you have a written implementation plan and no sub-agent tools (or you choose inline execution), the agent works through the plan's tasks itself in the current session, following test-first steps, checking each step's expected output, and recording progress in a ledger file so it can resume after context is compressed. It runs a final whole-branch review, ideally with a fresh-context reviewer, and only stops to ask on destructive actions, security matters, outside-the-worktree side effects like merging or publishing, or a plan too broken to proceed.

Great for

  • Developers who already have a detailed plan from a planning step and want the agent to execute it in one session
  • Teams that want to save cost by avoiding a fresh implementer and reviewer per task
  • Anyone running long multi-task plans who needs progress tracked in a ledger file

Skip it if

  • Work where you want a review gate after every task, since subagent-driven development is better suited (the skill itself recommends that route)
  • Plans with vague or missing specs, because the agent makes many judgement calls ('rulings') on its own and those can be wrong
  • Repositories where the agent should not create worktrees, ledger files, or run cleanup commands such as deleting the plan's workspace

Try asking

  • “Execute the implementation plan in docs/plan.md inline in this session”
  • “Continue the plan from where the progress ledger left off after the context was compacted”
  • “Run the final whole-branch review on this feature branch and triage the findings”
Step-by-step guide: how to use Executing Plans InlineInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    executing-plans.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip executing-plans.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip executing-plans.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Write a complete plan with explicit Expected lines and a spec before handing it over, since the agent relies on them to check its work
  • Review the ledger file for the Ruling lines the agent recorded, especially any that change the plan
  • Keep a sub-agent reviewer available for the final review; the skill says self-review is weaker

What's inside

  • SKILL.md

Redistributed unmodified from jnMetaCode/superpowers-zh under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.