When to Use This Skill
- Use when
crossframe-suiteroutes an explicit CrossFrame task about propositions, debate, hidden premises, rebuttals, strongest opposing arguments, evidence requirements, or withdrawal conditions. - Use to test claims before they become strong judgments.
- Do not use independently unless the user explicitly names this sibling skill.
Packaged Source Note
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.
Limitations
- The skill body is intentionally Chinese-canonical; English metadata is for discovery and does not replace the original Chinese terms.
- Use only after explicit CrossFrame invocation or
crossframe-suiterouting; do not apply it as a generic default reasoning layer. - It structures analysis, drafting, and review, but does not replace source verification, domain expertise, or legal, medical, or financial judgment.
This skill does not trigger independently. All CrossFrame tasks are uniformly dispatched from the
crossframe-suiteentry point. Users do not need to call this skill directly; the suite loads it automatically when needed according to routing rules.
If after proposition argumentation you need to write an article, public commentary, reading notes, or case documentation, first read ../crossframe-suite/SKILL.md for overall dispatch; this skill is only responsible for proposition decomposition, pro/con structure, evidence requirements, and retraction conditions.
crossframe-debate is crossframe a parallel lightweight entry point, used to decompose a proposition into testable arguments, rather than to help one side win a debate.
Chinese is the authoritative semantics; English is only used for skill ids, file names, and interface descriptions. When there is a conflict in understanding between Chinese and English, Chinese terminology and Chinese judgment shall prevail.
Lightweight Entry Rules
After each trigger, first read the canonical skill and routing map, and do not copy the full CrossFrame text:
- Read
../crossframe/SKILL.md。 - Read
../crossframe/references/read-routing-map.md。 - If the proposition triggers high responsibility, public institutions, intimate relationships, long-term evolution, framework governance, AI reality verification, weak signals/opacity, inability to exit, instrumentalization, metaphor/source transparency, or article output, you must additionally read
../crossframe/references/continuity-bundles.md, and use as needed../crossframe/worksheets/source-continuity-check.md; if the joint reading is not completed, you may only downgrade. - Reuse
../crossframe/templates/read-state-capsule.mdthev5-read-state-capsulespecified by../crossframe/worksheets/source-anchor-integrity-check.md, and execute in scenarios of high responsibility, public, AI/procedural artifacts, lifecycle, subjects unable to exit, or article output../crossframe/SKILL.mdto fill it in; this skill does not reinvent source routing. - Read this directory's
protocols/debate-protocol.md。 - Read
templates/debate-analysis-output.md。 - Read as needed
references/debate-quality-gates.mdandreferences/debate-failure-patterns.md。
Append canonical routing according to proposition type:
- Public issues, platform governance, policy, institutions, real persons or organizations: follow the routing map into public institutions, strong judgment, high responsibility, evidence cost, or proposition verification materials; when involving the latest facts, you must check sources or downgrade.
- Relationships, family, care, boundaries, interpretive labor: follow the routing map into materials related to intimate relationships, healing transfer, chains of responsibility, or love/open action.
- Philosophy, meaning, first cause, nihilism, value propositions: follow the routing map into materials on conceptual explanation, open assertions, and framework boundaries.
- Sanctions, reputation, rights, qualifications, resources, public accusations: you must enter proposition verification and high-responsibility routing; if verification is not completed, it can only serve as an open assertion or a proposition pending verification.
Default Task
After receiving a proposition, output by default:
- Proposition rewriting: rewrite slogans, emotions, or value statements into testable propositions.
- Pro structure: the best version of the pro side, without strawmanning.
- Con structure: the best version of the con side, without writing the con side as stupid or malicious.
- Hidden premises: factual premises, causal premises, value premises, scale premises, responsibility premises.
- Evidence requirements: current evidence, missing evidence, high-cost evidence, unavailable evidence.
- Strongest rebuttal: the strongest objection each side must face.
- Reverse conditions: which facts, when they appear, require the original proposition to be downgraded, rewritten, or redirected.
- Retraction conditions: which evidence is sufficient to retract this judgment.
- More stable expression: rewrite strong conclusions into open assertions, conditional judgments, or propositions pending verification.
Workflow
- Define the proposition object: object, scale, time window, affected objects, judgment tier.
- Determine the proposition type: public issue, relational proposition, organizational proposition, philosophical proposition, strong judgment, expressive rhetoric, or mixed proposition.
- Extract what is to be proven: what exactly does this proposition need to prove, and which parts are merely emotion, value preference, or rhetoric.
- Generate the best versions of both sides: steelman first, then criticize; straw men are not allowed.
- List the hidden premises: factual, causal, value, scale, responsibility-chain, and operability premises.
- Set the evidence threshold: what material can support, weaken, overturn, or fail to prove this proposition.
- Write the reversal conditions and retraction conditions: a proposition without retraction conditions cannot serve as a qualified conclusion.
- Output a more robust formulation: make the conclusion testable, downgradable, and modifiable by new facts.
Hard rules
- Do not write debate as mobilization, humiliation, camp labels, or character judgment.
- Do not use the weakest opposing side to prove your own side correct.
- Do not advance unilaterally: even if the user specifies a stance, point out the evidence and rebuttals that stance fears most.
- Do not treat anger, hurt, a sense of justice, disgust, or resonance as the argument itself.
- Do not output strong judgments without retraction conditions.
- Do not use grand scales to wash away low-scale suffering, responsibility chains, evidence gaps, or action boundaries.
- Do not treat AI reports, self-assessments, institutional statements, apology drafts, popularity, or polished expression as high-cost evidence.
- Do not treat CrossFrame terminology as a conclusion; terminology can only help examine structure.
Default output
Use by default templates/debate-analysis-output.md. If the user only wants a short answer, the minimal structure must still be retained:
- Proposition tier
- Best versions of both sides
- Most critical hidden premise
- Strongest rebuttal
- Retraction conditions
- More robust formulation
Qualification self-check
Check before output:
- Has this proposition already changed from a slogan into a testable statement?
- Have both sides been written as their best versions, rather than one side being demonized?
- Have factual, value, causal, scale, and responsibility premises been distinguished?
- Has it been explained what evidence is needed and what evidence is insufficient?
- Have reversal conditions and retraction conditions been given?
- Can the more robust formulation still retain the edge of the original question without overstepping the evidence?