structured-autonomy-plan Skill (awesome-copilot) for Claude Code: What It Does and How to Install It
Makes Claude research a feature request and write a commit-by-commit plan for one pull request before any code is written.
- Installs
- 8.8K
- on skills.sh
- Stars
- 40.7K
- on GitHub
- Updated
- Oct 2026
- last repo commit
This skill makes Claude plan before it codes. It researches the feature, splits the work into steps that each match one commit in a single pull request, and saves the plan to plans/{feature-name}/plan.md. It leaves [NEEDS CLARIFICATION] markers, asks you questions and waits for your feedback.
The folder has one file, SKILL.md, and no scripts. The research step calls a #tool:runSubagent tool, with a fallback to doing the research itself. The plan template names ResizeMe in one line, so point Claude at your own project.
Good for
- Developers who want a reviewed plan before an agent writes code
- Teams splitting one feature into testable commits
- Anyone who wants open questions flagged in the plan
Install structured-autonomy-plan
Claude Code
- 1
Download the skill folder from https://github.com/github/awesome-copilot/tree/HEAD/skills/structured-autonomy-plan.
- 2
Put it at ~/.claude/skills/structured-autonomy-plan/ so that SKILL.md sits directly inside.
- 3
Start a new session; the skill loads when its description matches the task.
Pick one route. The Copilot CLI plugin route installs the structured-autonomy plugin from the awesome-copilot marketplace: 3 skills (structured-autonomy-generate, structured-autonomy-implement, structured-autonomy-plan) and no agents. The other routes install only structured-autonomy-plan. The repo README says its customizations come from third-party developers, so inspect them before installing.
Built from the package data we hold. Not yet tested by us.
What's in the skill
When it loads
Structured Autonomy Planning Prompt
Sections in SKILL.md
- Step 1: Research and Gather Context
- Step 2: Determine Commits
- Step 3: Plan Generation
- Goal
- Implementation Steps
- Step 1: {Step Name} [SIMPLE features have only this step]
- Step 2: {Step Name} [COMPLEX features continue]
- Step 3: {Step Name}
Key facts
- License
- MIT (repo LICENSE)
- Installs on skills.sh
- 8,844 (Nov 2, 2026)
- Last repo commit
- Oct 30, 2026
- Audits on skills.sh
- Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass
- What it includes
- 1 file: SKILL.md (no scripts)
What it includes
Runs a sub-agent to gather context, then stops further tool calls
The research guide covers code context, documentation, dependencies (it uses #context7 if available) and existing patterns. It stops at 80% confidence.
Try: “Plan a feature that adds hotkey presets to my app”
Network ·
Step 1: ResearchOne commit for a simple feature, several testable commits for a complex one
The whole plan is assumed to land in a single pull request on its own branch.
Read ·
Step 2: Determine commitsSaves plans/{feature-name}/plan.md with NEEDS CLARIFICATION markers
The template has a title, branch name, description, goal and steps. Each step lists files, what changes and how to test it.
Write ·
Step 3: Write the planAsks clarifying questions, pauses for your feedback, revises the plan
SKILL.md says the pause is mandatory. After feedback it goes back to research if needed.
Read ·
Ask and revise
From the official docs
Permissions
Low risk
What it can do
- Read your data
- Create and change things
- Delete things
- Read-only mode
Watch out for
It writes a plan file in your repo
Claude saves plans/{feature-name}/plan.md and then waits for your feedback. SKILL.md says it writes no code in this step. github.com
The plugin route adds two more skills
The structured-autonomy plugin lists 3 skills and no agents. Use a folder route if you want only the planning skill. github.com
Technical details
| Turning on read-only | No read-only mode. |
|---|---|
| Sign-in and keys | No credentials are named in SKILL.md. |
| Scopes |
|
Questions
What does the structured-autonomy-plan skill do?
It makes Claude research a feature request and write a plan file, split into steps that match individual commits in one pull request. SKILL.md says it writes no code in this step.
Where does the plan go?
SKILL.md says to save it as plans/{feature-name}/plan.md and then ask your clarifying questions before going on.
Are there companion skills?
Yes, in the same plugin. structured-autonomy-generate takes a finished plan.md and writes an implementation.md, and structured-autonomy-implement carries out an implementation plan. You can use this one on its own.
Does it need a special tool?
SKILL.md names a #tool:runSubagent tool for the research step. If it is unavailable, SKILL.md tells the agent to run the research guide itself with its other tools.
How do I install it?
Use the install box at the top of this page. The repo README documents Copilot CLI plugins from the awesome-copilot marketplace, and the structured-autonomy plugin includes this skill. The box also has folder routes that install only this skill.
Why does the plan mention ResizeMe?
One line of the research guide says to identify how similar features are implemented in ResizeMe, a specific project. Our advice: tell Claude to use your own project instead.
Is it safe?
It writes a plan file in your repo and asks you for feedback before going on. Review the plan before you hand it to an implementation step. The plugin route also installs two more skills.
Information
updated Nov 2, 2026
- Repository
- github/awesome-copilot
- Folder
skills/structured-autonomy-plan
Sources
Docs, issue trackers and discussions this page is built from.
Show all 6Hide
- structured-autonomy-plan SKILL.md · github.com
- awesome-copilot README · github.com
- structured-autonomy plugin.json · github.com
- awesome-copilot LICENSE · github.com
- structured-autonomy-generate SKILL.md · github.com
- structured-autonomy-implement SKILL.md · github.com