Task Breakdown

The reason "launch the product" doesn't get done is that it's not a task — it's a project, masquerading as a task on a list. The first action is unclear, so nothing happens. This skill decomposes the goal into work streams (independent enough to run in parallel), then into MECE tasks (mutually exclusive, collectively exhaustive, estimable), then into a dependency map. The rule is simple: if you can't estimate a task with reasonable confidence, you haven't broken it down enough. A 4-hour task can be estimated. A 40-hour task is a guess wearing a number.

Then it finds the critical path — the longest chain of dependent work — because that's what actually determines the minimum project duration. Most people optimise the wrong tasks. They pick whatever's easiest or most fun, not whatever's on the critical path. The skill names the critical path explicitly so you start there. Parallel work (tasks not on the critical path) becomes your "fill the slack" pile. Realistic timelines are calculated by multiplying the critical path duration by 1.3-1.5 — that's not pessimism, it's accounting for the tasks you haven't thought of yet.

It handles the patterns where breakdown isn't enough. If the project is stalled, the skill diagnoses before re-breaking-down: usually the first task is still too big (next action should be doable in under 30 minutes), or the critical path is blocked, or there's a missing piece of information that needs to become a task on its own ("get the answer to X"). If the goal is wildly disproportionate to the timeline ("build a startup in two weeks"), it refuses to break down the impossible scope and instead helps define the most valuable subset that fits — an MVP or Phase 1. Then it breaks that down.

Mid-project re-estimation is built in. The original estimate is a hypothesis; once Phase 1 actuals exist, you have a calibration factor. If Phase 1 took 1.5x the estimate, apply that to remaining phases. Surface the revised timeline at the next milestone, not at the deadline — silent slippage destroys trust. For team capacity, the skill maps real hours (excluding meetings and other projects) and flags the bottleneck — usually one person on the critical path with the least time. The fix is reassign non-critical work, not "work harder."