closure-retrospectivelisted

# Closure Retrospective ## Core Principle Treat an observed outcome as evidence. Codify only the reusable mechanism that explains why a controllable decision caused, protected, or improved that outcome. Prefer the narrowest existing artifact and require the expected benefit to exceed its maintenance cost. This skill is not a memory workflow; do not propose or update memory unless the user explicitly requests it. ## Closure Gate Run the retrospective only after the substantive task is complete or clearly entering final handoff. If implementation, debugging, or research remains unresolved, stop the retrospective and finish the task first. Drop trivial work, one-off preferences, temporary debugging notes, and unsupported interpretations. ## Workflow 1. **Gather evidence.** Record observable behavior from the current task and the resulting failure, success, protected boundary, or efficiency gain. Keep observation separate from interpretation. Complete this step when every candidate names its evidence and outcome. 2. **Locate the decision point.** Identify a choice, sequence, validation, or scope decision available before the outcome. Complete this step when the future trigger is observable before the outcome and the alternative action is under the agent's control. Drop candidates with no such decision point. 3. **Extract the reusable mechanism.** Explain the supported relationship between the decision and outcome. For every candidate reaching this step, read [references/me