SuperAPI AI Directory
Partner中文AI newsFavorites
Skillscrossframe-public

crossframe-public

Writing & research

View repository

Updated 10-16 08:39

About this skill

CrossFrame Public

When to Use This Skill

  • Use when crossframe-suite routes an explicit CrossFrame task about public issues, platform governance, policy, institutional responsibility, public commitments, appeals, or compliance materials.
  • Use when source ledgers, evidence downgrades, public responsibility boundaries, and low-power subject protection matter.
  • 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-suite routing; 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-suite entry point. Users do not need to call this skill directly; the suite loads it automatically when needed according to routing rules.

If, after public issue analysis, you need to write a commentary article, organizational recommendations, debate arguments, or a quality review, first read ../crossframe-suite/SKILL.md for overall dispatch; this skill is only responsible for public facts, evidence boundaries, and specialized judgments on procedure and institutions.

CrossFrame Public is the crossframe specialized lightweight entry point for public issue/institutional commentary, and does not copy the full canonical CrossFrame text. Chinese is the authoritative semantics; English is only used as a skill id, file name, or external brief description.

Must read

After each trigger, first read:

  1. ../crossframe/SKILL.md
  2. ../crossframe/references/read-routing-map.md
  3. If the public judgment triggers high responsibility, public institutions, long-term evolution, framework governance, AI reality verification, weak signals/opacity, inability to exit, instrumentalization, metaphor/source transparency, or article output, 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.
  4. Reuse ../crossframe/templates/read-state-capsule.md as stipulated in v5-read-state-capsule, and execute in high-responsibility, public, AI/procedural artifact, lifecycle, non-exitable subject, or article output scenarios ../crossframe/worksheets/source-anchor-integrity-check.md. If the capsule is missing, return to ../crossframe/SKILL.md to fill it in; this skill does not reinvent source routing.
  5. protocols/public-issue-protocol.md
  6. references/source-and-evidence-rules.md
  7. ../crossframe/references/source-ledger-workflow.md, used to uniformly record source, time, source type, supported propositions, what cannot be proven, evidence tier, place of use, reasons for downgrading, and points still requiring further evidence.

Public commentary, platform governance, institutional compliance, and strong public judgments trigger by default v5-public-power-institution-pack、v5-low-power-protection-pack、v5-evidence-downgrade-action-ceiling-pack; AI reports or compliance materials additionally trigger v5-ai-process-artifact-boundary-pack。

Add according to task type:

  • Platform penalties, bans, throttling, post deletion, account appeals: read protocols/platform-appeal-protocol.md and templates/action-boundary.md。
  • Public policy, institutional commentary, fulfillment of public commitments: read protocols/public-policy-protocol.md and templates/public-comment-draft.md。
  • Institutional self-inspection, rectification reports, AI compliance materials, ethics/safety statements: read protocols/institutional-compliance-protocol.md and references/ai-compliance-performance.md。
  • Need to write a public commentary article: also read ../crossframe-essay/SKILL.md, but factual boundaries and evidence tiers still use this skill as the entry point.
  • Only boundaries required, not commentary: use templates/evidence-boundary-summary.md or templates/action-boundary.md。

Default source checking

Real public issues require source checking by default, and build a source ledger according to ../crossframe/references/source-ledger-workflow.md . Prioritize original materials, official texts, platform rules, original policy texts, regulatory/judicial/audit documents, first-hand statements by involved parties, cross-reports by credible media, and verifiable data.

If the user explicitly prohibits internet access or source checking is currently impossible:

  • Do not output strong judgments.
  • Do not treat popularity, retellings, screenshots, platform statements, or institutional self-assessments as facts.
  • Output 证据边界摘要 or 行动边界, and mark it as "no source checking, can only serve as a framework pending verification"; if user materials already exist, also state clearly what these materials can support and what they cannot prove.

Core checks

Public issue output must check five groups of questions:

  • Procedural justice: whether rules are publicly disclosed in advance, whether application is consistent, whether evidence is visible, and whether review is independent.
  • Appeal effectiveness: whether the appeal entry point is reachable, whether reasons can be submitted, whether replies are specific, and whether error correction truly changes the outcome.
  • Weak signal protection: whether complaints, anomalous data, minority testimony, and harm to marginalized groups are drowned out by popularity or institutional rhetoric.
  • Public commitment fulfillment: whether apologies, rectification, compensation, and commitments translate into verifiable resources, deadlines, responsible parties, and feedback mechanisms.
  • AI compliance performance risk: whether polished reports, self-assessment checklists, model-generated materials, and ethical slogans replace external verification and real constraints.

Evidence tiers

Before output, classify the material into:

  • Verified facts: can be supported by original text, records, reviewable data, or multi-source cross-checking.
  • High-cost evidence: material that carries legal, organizational, economic, reputational, or operational costs.
  • Low-cost claims: platform announcements, institutional self-assessments, PR copy, apologies without details, AI-generated compliance text.
  • Weak signals: early signals that have not yet formed a conclusion but point to harm, dysfunction, suppression, or abnormality.
  • Heat signals: search volume, reposts, comments, topic rankings; they can only indicate attention, not directly indicate truth or falsity.
  • Explanation/judgment: open assertions or commentary judgments formed based on facts and candidate mechanisms.

Output modes

Choose one main output according to user intent:

  • Public institutional diagnosis: explain the institutional object, factual boundaries, procedures/appeals/weak signals/commitment fulfillment/AI compliance risk, and candidate mechanisms.
  • Public commentary draft: first give the evidence boundaries and central proposition, then write a publishable commentary draft.
  • Evidence boundary summary: list verified, unverified, low-cost claims, heat signals, counterconditions, and next-step verification.
  • Action boundaries: give low-risk, retractable, recordable, and reviewable action suggestions; does not replace legal, medical, safety, or professional advice.

Hard rules

  • When sources have not been checked, do not pretend they have been checked; only downgrade the tier.
  • Do not treat heat as fact, and do not treat platform/institutional statements as strong evidence.
  • Do not omit the "what cannot be proven" and "reasons for downgrading" in the source ledger.
  • Do not write public issues as personality trials, moral condemnations, camp labels, or humiliation mobilization.
  • Do not use CrossFrame terminology to replace evidence verification, professional domain knowledge, or legal judgment.
  • Do not write "compliance materials exist" as "compliance has occurred".
  • Do not hide evidence gaps, counterconditions, or material that could lead to retracting a judgment for the sake of sharp commentary.
  • When real people, organizations, rights, sanctions, qualifications, or public memory are involved, follow ../crossframe/references/read-routing-map.md to enter the high-responsibility/proposition verification/public institution-related routing.

Minimum qualified output

A qualified output must at least answer:

  • What is the public object of this discussion?
  • Which facts have been verified, and which are only claims, heat, or explanations?
  • Are procedural justice and the effectiveness of appeals visible?
  • Who bears the cost, and who has the power to change the conditions?
  • Are weak signals protected, or are they drowned out by heat/rhetoric?
  • Do public commitments have a fulfillment path?
  • Is there an AI compliance performance risk?
  • What tier is this judgment at, and what evidence would cause it to be retracted or escalated?
  • What should be checked, said, and done next, and what cannot be done?