autoplan

Execute sequential CEO, design, eng, and DX reviews with auto-decisions.

Updated Apr 13, 2026
One-click install
npx skills add https://github.com/mgher668/surf-ai --skill autoplan-mgher668
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: autoplan
Source: https://github.com/mgher668/surf-ai/tree/main/.agents/skills/gstack/autoplan
Command: npx skills add https://github.com/mgher668/surf-ai --skill autoplan-mgher668

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinates an end-to-end review pipeline by reading multiple skill components (CEO, design, eng, and DX) from disk and executing them sequentially with auto-decisions based on six principles.

Core Features & Use Cases

  • Orchestrates CEO, design, eng, and DX reviews sequentially using six decision principles.
  • Surfaces taste decisions (close approaches, borderline scope, codex disagreements) at a final approval gate.
  • One-command run to produce a fully reviewed plan for stakeholders and teams.

Quick Start

Run /autoplan to execute the full review pipeline and generate a complete plan.

Frequently Asked Questions about autoplan

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate an end-to-end plan review pipeline?

Automating an end-to-end plan review pipeline involves executing multiple review components sequentially with auto-decisions. This Skill coordinates CEO, design, eng, and DX reviews from disk to produce a fully reviewed plan via a single command.

What is the best way to run sequential design and engineering reviews for a project plan?

Running sequential design and engineering reviews requires orchestrating multiple perspectives with deterministic control flow. This process auto-decides based on six principles and surfaces borderline scope or taste decisions at a final approval gate.

How does auto-decision orchestration work for multi-stage plan reviews?

Auto-decision orchestration in plan reviews works by reading skill components from disk and executing them sequentially using six decision principles. It handles errors robustly and outputs an artifact for downstream planning dashboards.

Can I surface borderline scope and taste decisions at a final approval gate?

Surfacing taste decisions at a final approval gate isolates close approaches, borderline scope, and codex disagreements for stakeholder review. The pipeline executes all automated stages first, then pauses for human input on these subjective points.

Do I need to prepare specific input files to run a full plan review pipeline?

Running a full plan review pipeline requires having the necessary skill components available on disk. The orchestration reads these components directly to execute the sequential auto-decisions and generate the final reviewed plan artifact.

What are the limitations of using a single command to auto-review a full project plan?

Using a single command to auto-review a full plan limits granular control over individual review stages. While it provides deterministic sequential execution and robust error handling, subjective taste decisions still require manual resolution at the final gate.