10x-plan

Create detailed implementation plans through iterative research and structured questioning.

Updated May 20, 2026
One-click install
npx skills add https://github.com/Oskarovsky/Sprinter --skill 10x-plan-oskarovsky
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: 10x-plan
Source: https://github.com/Oskarovsky/Sprinter/tree/main/.cursor/skills/10x-plan
Command: npx skills add https://github.com/Oskarovsky/Sprinter --skill 10x-plan-oskarovsky

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Turning a vague task or ticket into an actionable implementation plan requires deep codebase research and many decisions; this Skill guides an interactive, iterative process that produces thorough technical specifications without costly rework. ## Core Features & Use Cases - Adaptive Questioning: Scales the number and focus of clarifying questions based on task complexity (LOW/MEDIUM/HIGH) and upstream artifacts like frame briefs and research docs. - Parallel Codebase Research: Spawns Explore and general-purpose sub-agents to map files, trace data flow, and surface conventions before writing the plan. - Structured Plan Output: Writes a phased plan to context/changes/<change-id>/plan.md with success criteria, a machine-readable Progress section, and a two-page plan brief. - Use Case: Given a ticket to add OAuth login, the Skill reads prior research, asks targeted architecture and edge-case questions, then produces a phased plan with automated and manual verification steps. ## Quick Start Invoke the skill with a change id or task description, for example by asking to create an implementation plan for the oauth-login change.

Frequently Asked Questions about 10x-plan

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

FAQPage Schema
How do I create a detailed implementation plan from a ticket?

Invoke the skill with the task description, ticket reference, or a change id. It reads referenced files, researches the codebase with parallel sub-agents, asks structured clarifying questions, then writes a phased plan to context/changes/<change-id>/plan.md.

How does the skill decide how many questions to ask?

It assesses task complexity as LOW (4-6 questions), MEDIUM (7-10), or HIGH (11-15), then reduces the count when upstream artifacts exist. A frame brief or research doc means fewer questions because decisions already made are not re-asked.

Can I use a frame brief or research doc as input to planning?

Yes. Passing context/changes/<change-id>/frame.md or research.md scales down questioning: the frame's problem statement is treated as authoritative and the research doc's code references become the codebase baseline.

What files does the implementation plan produce?

It writes plan.md with phases, success criteria, and a Progress checklist section, plus plan-brief.md, a two-page summary of what, why, and key decisions. It also updates change.md status to planned.

When should I not use this planning approach?

Skip it for trivial single-file changes where full questioning adds no value, and do not use it on archived changes. If the resolved path is under context/archive/, the skill refuses and tells you to open a new change instead.