What problem does it solve? After a development session that used the kiro spec-driven workflow, there is no structured way to capture what actually worked, what wasted time, and which parts of the framework need fixing. This Skill turns a finished session into two dated, comparable reports: an engineer-experience feedback report and a triaged skill-improvement backlog for the framework maintainer. ## Core Features & Use Cases - Bounded developer interview: Runs a fixed-core interview (reuse intent, what it caught, what wasted time) with a strict question budget, logging every question and answer verbatim. - Evidence-chained friction analysis: Every reported friction must cite a trigger, signal, and cost anchored to spec artifacts, gate verdicts, or in-session corrections, with near-misses flagged as latent risks. - Triage-filtered improvement backlog: Routes each finding to SKILL-FIX, PROCESS, one-off, or LEARNING buckets so feedback does not translate 1:1 into skill edits. - Use Case: At the end of a feature implementation driven by /kiro:* commands, ask for a retro to produce feedback-<ts>.md and skill-improvements-<ts>.md under .kiro/, giving the maintainer comparable session rows over time. ## Quick Start Ask the agent to retro this session and generate the feedback and skill-improvement reports for the feature you just completed.
