plan-large-work

Plan multi-session engineering work as a map of investigation tickets resolved one per session.

1|Updated Jul 13, 2026
One-click install
npx skills add https://github.com/dineshrevunuru/SuperSkills --skill plan-large-work-dineshrevunuru
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan-large-work
Source: https://github.com/dineshrevunuru/SuperSkills/tree/main/plan-large-work
Command: npx skills add https://github.com/dineshrevunuru/SuperSkills --skill plan-large-work-dineshrevunuru

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Large migrations, refactors, and specs that exceed one agent session tempt linear plans that commit a build order before key unknowns are resolved, collapsing when the first unknown surfaces. This Skill charts the unknowns as a map of investigation tickets and resolves them one per session until the build order is clear. ## Core Features & Use Cases - Destination-first charting: Name the destination, sketch the fog breadth-first, and cut only the tickets sharp enough to specify, leaving the rest as Not-yet-specified fog. - Solo MAP.md default: Track tickets, blocking edges, the frontier, and decisions in a local markdown file, with a mode-switch up to GitHub Issues or Linear only when a real tracker exists. - Typed tickets with routing: Research, prototype, decision, and task tickets each route to the appropriate downstream skill, with one ticket resolved per session and a handoff at the end. - Use Case: For a Pages-to-App-Router migration, chart research tickets on auth and data-fetching first, block the order-committing decision behind them, and graduate fog into tickets as answers sharpen. ## Quick Start Ask the agent to map out the migration or refactor as investigation tickets in a MAP.md file, resolving one frontier ticket per session.

Frequently Asked Questions about plan-large-work

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

FAQPage Schema
How do I plan a large code migration across multiple sessions?

Chart the unknowns as investigation tickets in a MAP.md file instead of writing a linear plan. Name the destination first, sketch the fog breadth-first, cut the sharp tickets, then resolve one frontier ticket per session until the build order is clear.

What is the difference between plan-large-work and plan-work-tickets?

plan-large-work resolves the unknowns that gate the build order before any committed spec exists, producing decisions. plan-work-tickets slices an already-committed plan into buildable tickets that ship code; the map's committed destination is its input.

Can I use GitHub Issues or Linear instead of a local MAP.md file?

Yes, but only when a real tracker already exists. The solo default is a local MAP.md with Blocked-by title lines; the tracker path uses a map-labelled issue, child tickets, native dependency edges, and claiming by assignment.

When should I not use an investigation map for planning?

Skip the map when the work fits one or two sessions with no real fog, when framing the problem itself, or when slicing an already-committed spec into build tickets. A breadth-first sketch that surfaces no unknowns means you can plan inline.

Why should I resolve only one ticket per session?

One ticket per session keeps each resolution focused and forces a clean handoff, so the map stays an accurate cross-session index. Resolving several tickets at once blurs decisions and loses the frontier state the next session depends on.