What problem does it solve? After a release ships, teams often move on without ever checking whether the feature met its goals, so the next planning cycle starts from memory instead of evidence. This Skill runs the post-release learning checkpoint: it scores what shipped against the PRD's committed success criteria, extracts insights, forces an explicit decision per criterion, and updates the PRD so the next cycle starts from facts. ## Core Features & Use Cases - Success-criteria scorecard: Builds a target-vs-actual-vs-verdict table for every criterion written before launch, flagging unmeasurable criteria as owned instrumentation gaps. - Multi-channel evidence gathering: Pulls product metrics, error trends via sentry-cli, business impact via SaaS metrics skills, and qualitative feedback before forming any opinion. - Explicit decision tagging: Every verdict gets a persevere, iterate, expand, or kill decision with an owner, routing kills to eol-message and iterates back into prd-development. - Use Case: A feature shipped six weeks ago and its checkpoint date arrives. Run this Skill to compare activation and retention actuals against the PRD targets, discover one criterion was never instrumented, decide to iterate on onboarding, and append the full review to the PRD. ## Quick Start Run the learn-iterate checkpoint for the checkout-redesign release that shipped last month and update its PRD with the scorecard and decisions.
