answer-contract is an output-discipline skill. The public repo gives you the four rules and worked cases. The execution layer behind them is not published — it is delivered through collaboration.
- Task-class implied-need checklists — the full "what did the request actually imply" checklists for code tasks, marketing copy, and decision memos, with the confirm-template and veto-flow.
- Option-matrix templates — the dimension sets and scoring rubrics used to build 3-Option-Plus-Premium comparisons in your domain.
- Calibration data — before/after calibration from live usage, plus the pre-send check as an automated gate rather than a manual ritual.
- Vertical adaptation packs — the four rules adapted to specific domains (e-commerce content, developer tooling, knowledge work), with vocabulary and kill-signal presets.
- Teams shipping AI assistants or coding agents whose output quality is user-facing.
- Tooling vendors who want output-contract discipline embedded in their product.
- Internal platform teams rolling out shared standards for LLM-assisted work.
- It will not claim capabilities beyond the rules (this skill constrains output shape and decision hygiene; it does not improve raw model capability).
- It will not publish calibration data or client-derived material.
Email: hcac4735@agent.qq.com — include "four-rules" in the subject, describe your use case and what you want unlocked. Replies within 48h.