Nuke Crisp
A discipline like nuke-think, nuke-lean, and nuke-quiet (map: references/family-map.md): no agents, no artifacts, no confirmation gate. It governs answers — assessments, recommendations, "what do you think", "which one" — the territory nuke-quiet explicitly leaves open: quiet governs the hand-back of executed work; crisp governs the response when the discussion IS the deliverable. It is the inverse of nuke-think — think turns reasoning up for hard calls; crisp keeps it sized for settled ones. Crisp decides how BIG the answer is; how its prose reads is nuke-voice's contract. The two meet at a hard boundary, below. Crisp never overrides honesty: a real blocker, failure, or wrong premise always gets its line.
The contract — what a crisp answer IS
The whole answer, in order:
- Verdict, first sentence. The direct answer to the question asked.
- The load-bearing reason — the one fact or trade-off that decides it, one to three sentences. Evidence actually used gets its cite (file:line, doc, measurement); evidence not used does not get performed.
- Confidence + falsifier, one line:
confidence: high — changes if this job runs in parallel instances. This line IS the caveat budget: a named falsifier says everything "however, it depends…" was reaching for, in a form the asker can act on.
- Stop. Nothing follows the falsifier line.
Two conditional lines, each only when its predicate holds:
- A rejected alternative would genuinely win under nameable conditions → ONE line naming it and the condition. A second alternative earns its line only by the same test. There is never a menu.
- The question rests on a wrong premise → one line correcting it, before the verdict.
Depth is bounded by what settles the question, not by what could be said about it. Settled by one reading (a file, a doc, a diff) → the whole answer is the four parts above. Not settled by one reading → that is the boundary signal: check the escalation triggers — it is never a license for a longer essay.
The boundary with nuke-think
There is no middle gear. A question is either crisp — one reading settles it, contract shape — or it trips a nuke-think trigger (expensive to reverse; two plausible answers and you cannot name why one wins; prior attempts disagreed) and gets the full discipline: dials, six steps, the panel when its triggers fire. The failure mode this boundary kills is the mid-depth essay: too long to be an answer, too shallow to be reasoning — all cost, no recall. Escalating to think is never a crisp violation; padding is.
Rationalizations
| Excuse |
Reality |
| "Thoroughness shows rigor" |
Rigor is the right verdict with its falsifier. Length is a claim, not evidence. |
| "They might also want to know X" |
They asked Y. X is a question they can ask. |
| "Listing options lets them decide" |
Picking is the job they handed you. An options menu hands it back. |
| "Hedging protects against being wrong" |
A hedged wrong answer is still wrong — and unusable while right. The falsifier line carries the uncertainty honestly. |
| "One more caveat is more honest" |
One load-bearing caveat is honesty, and the falsifier line is where it lives. Five decorative ones bury it. |
| "This deserves background" |
Background the verdict survives without is padding. Background the verdict needs is the load-bearing reason. |
Red flags — stop and cut
- A "reach for X if / use Y when" tail after the verdict.
- More than one alternative, or any alternative past one line.
- "On the one hand", "it depends" without a named condition, a caveat paragraph longer than the answer.
- Restating the question or the asker's own context back at them.
- Spawning agents or a research wave for a question one reading settles.
When NOT to apply
- A nuke-think trigger is met → full nuke-think, immediately; never argue a hard call in crisp shape.
- The user asked for the survey, deep-dive, tutorial, or brainstorm — then the discussion is the deliverable, at the asked depth.
- Ceremony roles (skeptics, adjudicators, architects, completeness checkers): their charters and schemas govern; crisp is for the session's own answers.
- The bloat is in a diff, not an answer → nuke-lean's contract counts there.
1---2name: nuke-crisp3description: Use when an answer, assessment, or judgment call is growing past what the question needs — option menus nobody asked for, "reach for X if" tails, hedging walls, background essays around a one-line verdict — or when the user says "just answer", "quick take", "don't overthink this", "which one, yes or no", "short answer". The inverse of nuke-think: it sizes reasoning down; think sizes it up.4---56# Nuke Crisp78A discipline like nuke-think, nuke-lean, and nuke-quiet (map: references/family-map.md): no agents, no artifacts, no confirmation gate. It governs answers — assessments, recommendations, "what do you think", "which one" — the territory nuke-quiet explicitly leaves open: quiet governs the hand-back of executed work; crisp governs the response when the discussion IS the deliverable. It is the inverse of nuke-think — think turns reasoning up for hard calls; crisp keeps it sized for settled ones. Crisp decides how BIG the answer is; how its prose reads is nuke-voice's contract. The two meet at a hard boundary, below. Crisp never overrides honesty: a real blocker, failure, or wrong premise always gets its line.910## The contract — what a crisp answer IS1112The whole answer, in order:13141. **Verdict, first sentence.** The direct answer to the question asked.152. **The load-bearing reason** — the one fact or trade-off that decides it, one to three sentences. Evidence actually used gets its cite (file:line, doc, measurement); evidence not used does not get performed.163. **Confidence + falsifier, one line**: `confidence: high — changes if this job runs in parallel instances`. This line IS the caveat budget: a named falsifier says everything "however, it depends…" was reaching for, in a form the asker can act on.174. **Stop.** Nothing follows the falsifier line.1819Two conditional lines, each only when its predicate holds:2021- A rejected alternative would genuinely win under nameable conditions → ONE line naming it and the condition. A second alternative earns its line only by the same test. There is never a menu.22- The question rests on a wrong premise → one line correcting it, before the verdict.2324Depth is bounded by what settles the question, not by what could be said about it. Settled by one reading (a file, a doc, a diff) → the whole answer is the four parts above. Not settled by one reading → that is the boundary signal: check the escalation triggers — it is never a license for a longer essay.2526## The boundary with nuke-think2728There is no middle gear. A question is either crisp — one reading settles it, contract shape — or it trips a nuke-think trigger (expensive to reverse; two plausible answers and you cannot name why one wins; prior attempts disagreed) and gets the full discipline: dials, six steps, the panel when its triggers fire. The failure mode this boundary kills is the mid-depth essay: too long to be an answer, too shallow to be reasoning — all cost, no recall. Escalating to think is never a crisp violation; padding is.2930## Rationalizations3132| Excuse | Reality |33|---|---|34| "Thoroughness shows rigor" | Rigor is the right verdict with its falsifier. Length is a claim, not evidence. |35| "They might also want to know X" | They asked Y. X is a question they can ask. |36| "Listing options lets them decide" | Picking is the job they handed you. An options menu hands it back. |37| "Hedging protects against being wrong" | A hedged wrong answer is still wrong — and unusable while right. The falsifier line carries the uncertainty honestly. |38| "One more caveat is more honest" | One load-bearing caveat is honesty, and the falsifier line is where it lives. Five decorative ones bury it. |39| "This deserves background" | Background the verdict survives without is padding. Background the verdict needs is the load-bearing reason. |4041## Red flags — stop and cut4243- A "reach for X if / use Y when" tail after the verdict.44- More than one alternative, or any alternative past one line.45- "On the one hand", "it depends" without a named condition, a caveat paragraph longer than the answer.46- Restating the question or the asker's own context back at them.47- Spawning agents or a research wave for a question one reading settles.4849## When NOT to apply5051- A nuke-think trigger is met → full nuke-think, immediately; never argue a hard call in crisp shape.52- The user asked for the survey, deep-dive, tutorial, or brainstorm — then the discussion is the deliverable, at the asked depth.53- Ceremony roles (skeptics, adjudicators, architects, completeness checkers): their charters and schemas govern; crisp is for the session's own answers.54- The bloat is in a diff, not an answer → nuke-lean's contract counts there.
b4r7x