research-to-decisionlisted

# Research To Decision Answer one decision question with retained evidence. Keep research, inference, and the final decision distinct. ## Frame The Packet State before searching or experimenting: - question and decision owner; - why the answer matters now; - authorized scope and prohibited actions; - preferred source hierarchy; - comparison or success criteria; - stop condition and expected repository artifact. Do not broaden a research task into implementation authority. ## Gather Evidence Prefer primary and current sources. For repository experiments, record the exact commit, environment, inputs, commands, and artifact paths. Separate each conclusion into: - **Documented fact:** directly supported by a cited source; - **Observation:** reproduced in the named environment; - **Inference:** reasoned from facts or observations; - **Unknown:** material uncertainty that remains. Avoid long source dumps. Retain the smallest excerpt or result needed to make the conclusion auditable. ## Compare Against The Decision Criteria Evaluate credible options consistently. Record constraints, trade-offs, failure modes, reversibility, and what evidence would change the recommendation. A failed candidate is not proof that every alternative works. ## Retain The Result Write the result into the repository's existing research, feasibility, or decision location. Include: - scoped status; - executive conclusion; - method and source index; - evidence with fact/observation/inference labe