Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when CrossFrame Suite routes explicit Chinese proposition testing, debate analysis, hidden-premise review, rebuttal design, or withdrawal condition checks.
.claude/skills/sickn33-crossframe-debate/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-22 | ✗→✓ | ▲ Improved | 314% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 66% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 89% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 42% | 0% |
crossframe-suite routes an explicit CrossFrame task about propositions, debate, hidden premises, rebuttals, strongest opposing arguments, evidence requirements, or withdrawal conditions.This AAS-ready copy preserves the original CrossFrame skill body below. Chinese remains the canonical semantic layer; English metadata is only for discovery, installation, and repository review.
crossframe-suite routing; do not apply it as a generic default reasoning layer.> 本 skill 不独立触发。 所有 CrossFrame 任务统一从 crossframe-suite 入口调度。用户无需直接调用本 skill；suite 根据路由规则在需要时自动加载。
如果命题论证之后要写文章、公共评论、读书笔记或案例沉淀，先读取 ../crossframe-suite/SKILL.md 做总调度；本 skill 只负责命题拆解、正反结构、证据要求和撤回条件。
crossframe-debate 是 crossframe 的平行轻入口，用于把一个命题拆成可检验论证，而不是帮助某一方赢辩论。
中文为权威语义；英文只用于 skill id、文件名和接口说明。遇到中英文理解冲突时，以中文术语和中文判断为准。
每次触发后，先读取 canonical skill 和路由图，不复制 CrossFrame 全文：
../crossframe/SKILL.md。../crossframe/references/read-routing-map.md。../crossframe/references/continuity-bundles.md，并按需使用 ../crossframe/worksheets/source-continuity-check.md；未完成联读时只能降档。../crossframe/templates/read-state-capsule.md 规定的 v5-read-state-capsule，并在高责任、公共、AI/过程性产物、生命周期、无法退出主体或文章输出场景执行 ../crossframe/worksheets/source-anchor-integrity-check.md。如果胶囊缺失，回到 ../crossframe/SKILL.md 补齐；本 skill 不重新发明源路由。protocols/debate-protocol.md。templates/debate-analysis-output.md。references/debate-quality-gates.md 和 references/debate-failure-patterns.md。按命题类型追加 canonical 路由：
收到命题后，默认输出：
默认使用 templates/debate-analysis-output.md。若用户只要短答，也必须保留最小结构：
输出前检查：
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→fail | 15,853 | 18,608 | +17% | 1 | 1 | 0% | 2,379 | 4,693 | +97% | 0 | 0 | — |
case-05 | pass→fail | 12,171 | 13,249 | +9% | 1 | 1 | 0% | 2,081 | 3,688 | +77% | 0 | 0 | — |
case-22 | fail→pass | 6,133 | 14,752 | +141% | 1 | 1 | 0% | 954 | 3,947 | +314% | 0 | 0 | — |
case-01 | fail→fail | 18,362 | 21,444 | +17% | 1 | 1 | 0% | 3,214 | 5,264 | +64% | 0 | 0 | — |
case-02 | pass→pass | 10,361 | 12,368 | +19% | 1 | 1 | 0% | 1,708 | 3,869 | +127% | 0 | 0 | — |
case-03 | pass→pass | 13,786 | 16,915 | +23% | 1 | 1 | 0% | 2,286 | 4,541 | +99% | 0 | 0 | — |
case-06 | pass→pass | 12,268 | 21,018 | +71% | 1 | 1 | 0% | 2,071 | 4,725 | +128% | 0 | 0 | — |
case-07 | pass→pass | 11,337 | 10,093 | -11% | 1 | 1 | 0% | 1,838 | 3,338 | +82% | 0 | 0 | — |
case-08 | fail→fail | 14,030 | 17,760 | +27% | 1 | 1 | 0% | 2,394 | 4,475 | +87% | 0 | 0 | — |
case-09 | fail→pass | 16,470 | 13,447 | -18% | 1 | 1 | 0% | 2,540 | 3,621 | +43% | 0 | 0 | — |
case-10 | pass→pass | 13,123 | 19,813 | +51% | 1 | 1 | 0% | 2,289 | 4,627 | +102% | 0 | 0 | — |
case-11 | pass→pass | 13,066 | 12,704 | -3% | 1 | 1 | 0% | 2,337 | 3,841 | +64% | 0 | 0 | — |
case-12 | fail→pass | 15,870 | 16,527 | +4% | 1 | 1 | 0% | 2,690 | 4,459 | +66% | 0 | 0 | — |
case-13 | fail→pass | 13,745 | 15,560 | +13% | 1 | 1 | 0% | 2,345 | 4,435 | +89% | 0 | 0 | — |
case-14 | pass→pass | 11,561 | 13,646 | +18% | 1 | 1 | 0% | 1,949 | 3,825 | +96% | 0 | 0 | — |
case-15 | fail→fail | 4,267 | 4,152 | -3% | 1 | 1 | 0% | 804 | 2,317 | +188% | 0 | 0 | — |
case-16 | fail→fail | 11,140 | 12,945 | +16% | 1 | 1 | 0% | 1,972 | 3,755 | +90% | 0 | 0 | — |
case-17 | pass→pass | 10,705 | 19,154 | +79% | 1 | 1 | 0% | 1,854 | 4,541 | +145% | 0 | 0 | — |
case-18 | pass→pass | 10,171 | 17,457 | +72% | 1 | 1 | 0% | 1,739 | 4,375 | +152% | 0 | 0 | — |
case-19 | fail→pass | 17,723 | 18,661 | +5% | 1 | 1 | 0% | 3,345 | 4,756 | +42% | 0 | 0 | — |
case-20 | pass→pass | 22,222 | 25,494 | +15% | 1 | 1 | 0% | 3,625 | 5,359 | +48% | 0 | 0 | — |
case-21 | pass→pass | 12,775 | 15,755 | +23% | 1 | 1 | 0% | 1,941 | 3,983 | +105% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.