tradeofflisted

# Tradeoff ## Overview A decision fork produces two or more options, and the vocabulary for costing them already exists across this repo — `Scope-risk:` and `Confidence:` trailers, the `Blast Radius` clause, the SRP clause's "buys a review round, a merge ordering, a tracker entry, and a reviewer's context reload — each, every time". What is missing is a surface that puts them side by side at the moment of choosing. `Decision-Point Briefing` covers scope, risk and reversibility but fires only "when you stop to ask", so a fork the agent walks straight through is costed by nothing. This skill fills that gap and nothing else. It enumerates the options, scores each on four axes, and marks where every score came from. **Core principle:** a convention is not an option. When the lookup returns a precedent, that precedent is the answer and no table is built. ## When to Use - An implementation fork with two or more real candidates — a pair (hook vs skill, inline vs extracted) or a wider set (enforce in a hook / in a skill / in CI / in review only) - Deciding whether to act on a review finding, and whether it blocks Not for deciding whether a piece of work is worth starting — the Issue Review anchor's `Judgement` table (Need / Alternative / Urgency) owns that, and a second surface for it would only drift from the first. ## Process ### Step 1: Enumerate the options — three probes, then stop Take the options the user named, then run exactly these three probes. Three, then st