Legal Document Summarization logo

Legal Document Summarization

Community
THUYRan
legal-document-summarization

Trigger this skill when a structured summary is needed for judgments, rulings, mediation statements, arbitral awards, administrative penalty decisions, and other legal instruments. Typical trigger scenarios include, but are not limited to: - The user asks for a summary of a legal instrument, extraction of key points, or generation of a holding / ratio abstract - The user needs to convert multiple legal instruments into standardized summaries that are comparable and searchable - The user asks to extract case information, issues in dispute, legal authorities, or reasoning - The user needs summary materials for case studies, litigation analysis, or teaching preparation - The user asks to distinguish summary emphases across document types (judgment / ruling / mediation / arbitration / administrative penalty) Core value of this skill: On the basis of fidelity to the original text, convert lengthy legal instruments into concise, objective, structured summaries that highlight core content and avoid restating the full text or offering subjective evaluation.

Overview

PublisherTHUYRan
RepositoryLegal-Skills-Chinese
Skill namelegal-document-summarization
Stars
876
Forks
97
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by THUYRan on GitHub. Read the source before you install it.

Installation

Install the Legal Document Summarization AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/THUYRan/Legal-Skills-Chinese/tree/main/English/skills/legal-document-summarization
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/THUYRan/Legal-Skills-Chinese.git /tmp/Legal-Skills-Chinese
mkdir -p .claude/skills
cp -r /tmp/Legal-Skills-Chinese/English/skills/legal-document-summarization .claude/skills/legal-document-summarization
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Legal Document Summarization in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Legal Document Summarization on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Legal Document Summarization is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Chinese source (authoritative): ../../skills/legal-document-summarization/SKILL.md

Legal Document Summarization

Overview Table

ItemContent
Capability ID15
Capability NameLegal Document Summarization
Capability TypeInformation extraction and structuring
Core FunctionIdentify document type, extract elements, organize logic, and produce structured output for various legal instruments
InputFull text of judgments, rulings, mediation statements, arbitral awards, administrative penalty decisions, and similar legal instruments
OutputStructured summary customized by document type (case information, facts, issues, authorities, conclusions, reasoning)
Related CapabilitiesLegal document formatting (ensure normative output format); statutory retrieval (verify accuracy of legal authorities); legal reasoning (analyze reasoning logic)
Key PrinciplesFidelity to the original; objectivity and neutrality; highlight the core; distinguish by type

Legal Disclaimer

Important notices:

  1. The summary must be faithful to the original text and must not add facts, views, or conclusions not stated in the original.
  2. The summary must not contain subjective evaluative language (e.g., "the court wisely held," "the penalty was reasonable," "the defendant acted in bad faith").
  3. For scanned or image-based instruments, OCR may introduce errors; uncertainties should be marked in the summary.
  4. Summary conclusions reflect only what the instrument records and do not constitute legal opinions on the merits of the case.
  5. For cases still pending, the summary must not disclose non-public procedural information or trade secrets.

I. Core Concepts

1.1 Document Types and Summary Emphases

Different types of legal instruments have different core functions and information structures; summary emphases must adjust accordingly:

Document TypeCore FunctionSummary EmphasesKey Elements Distinguishing It from Other Types
Judgment (panjueshu)Resolve substantive disputesFact findings, application of law, reasons for decision, dispositionThe court's "findings" of fact (not parties' allegations)
Ruling (caidingshu)Resolve procedural mattersProcedural background, application / objection content, ruling authorities, procedural conclusionDoes not address substantive disputes; only procedural issues
Mediation statement (tiaojieshu)Confirm the parties' agreementMatters in dispute, content of the mediation agreement, performance arrangementsReflects mutual consent, not a coercive court adjudication
Arbitral award (zhongcai caijueshu)Final award by the arbitral tribunalArbitration claims, fact findings, tribunal opinions, award dispositionFinality of one award; based on an arbitration agreement
Administrative penalty decision (xingzheng chufa juedingshu)Administrative sanctionUnlawful facts, penalty basis, type and measure of penaltyUnilateral administrative act; subject to reconsideration and litigation

1.2 Summary vs. Abridgment vs. Evaluation

Original text ──────────────────────────────────────→

  Summary: Extract core elements, reorganize structurally, omit non-critical details
  
  Abridgment: Compress the original proportionally, retaining a slimmed version of all information points
  
  Evaluation: Add subjective judgments and value commentary (prohibited in summaries)

Essence of summarization: Not merely shortening the original, but identifying core information, removing irrelevant details, and reorganizing information by logical structure.

1.3 Six-Element Extraction Framework

All legal-instrument summaries should center on the following six elements:

┌─────────────────────────────────────────────────┐
│  1. Case information: who, when, where, which body issued it │
│  2. Factual background: what happened (facts relevant to the conclusion) │
│  3. Issues in dispute: core legal questions to be resolved │
│  4. Legal authorities: which norms were cited │
│  5. Disposition / result: how the matter was finally handled │
│  6. Reasoning: how the conclusion was derived from facts + norms │
└─────────────────────────────────────────────────┘

II. Complete Workflow

Phase One: Preparation and Identification

Step 1: Identify Document Type

First determine which type the legal instrument belongs to:

Judgment CriterionJudgmentRulingMediation StatementArbitral AwardAdministrative Penalty Decision
Title keywords"Civil Judgment," "Criminal Judgment," "Administrative Judgment""Civil Ruling," "Enforcement Ruling""Civil Mediation Statement," "Mediation Agreement""Arbitral Award""Administrative Penalty Decision"
Core functionSubstantive adjudicationProcedural dispositionConfirm agreementArbitral adjudicationAdministrative sanction
Issuing bodyCourtCourtCourt (confirming the parties' agreement)Arbitration commission / tribunalAdministrative agency

Judgment principle: Do not apply a format before identifying the document type; otherwise the true emphases of that document type are easily missed.

Step 2: Extract Case Information

Extract complete citation information from the caption or title section:

Extraction ItemSource LocationExample
Case nameCaptionMeijing Co. v. Ant Co. unfair competition dispute
Case / document numberBelow the title(2024) Jing 0108 Min Chu No. 1234
DateEnd of the instrument7 December 2024
Issuing bodyCaption / endBeijing Haidian District People's Court
PartiesCaptionPlaintiff: Meijing Co.; Defendant: Ant Co.

Requirement: This section should retain original wording as far as possible to support retrieval and citation.

Phase Two: Content Extraction

Step 3: Distill Key Facts or Procedural Background

According to document type, distill facts or procedural matters directly related to the conclusion:

Judgments / arbitral awards / administrative penalty decisions:

  • Distill facts related to the substantive dispute
  • Distinguish "facts alleged by the parties" from "facts found by the instrument"
  • Retain only key facts that affect the conclusion; omit irrelevant details

Rulings:

  • Distill procedural background (e.g., after the plaintiff sued, the defendant raised a jurisdictional objection)
  • Distill the content of the application or objection
  • Do not include facts of the substantive dispute

Mediation statements:

  • Distill matters in dispute and the background of reaching mediation
  • Main points of divergence between the parties
  • Key factors facilitating mediation (if mentioned)
Step 4: Summarize Issues in Dispute

Distill the disputes in the instrument into one or more clear legal questions:

Requirements:

  • Phrase as legal questions (answerable interrogatives)
  • Capture the problems the decision truly resolved
  • Avoid overly broad background descriptions
Poor Issue FramingGood Issue Framing
"The parties had a dispute""Whether Meijing Co.'s scraping of data from Ant's platform constitutes unfair competition"
"There is a controversy""Whether the defendant's jurisdictional objection is established"
"The compensation issue""Whether the plaintiff's claimed RMB 5 million in damages has factual and legal basis"
Step 5: Extract the Disposition or Result

Accurately summarize the instrument's final disposition:

Document TypeManner of Stating the ConclusionExample
JudgmentGrant / dismiss / confirm / modify"Judgment that the defendant cease the unfair competition at issue and compensate the plaintiff RMB 2 million for economic losses"
RulingGrant / dismiss / transfer / stay / terminate"Ruling dismissing the defendant's jurisdictional objection"
Mediation statementThe parties reached the following agreement"The parties voluntarily reached the following mediation agreement: 1. The defendant shall pay the plaintiff RMB XX by [date]..."
Arbitral awardAwards that..."Awards that the respondent pay the claimant RMB 1 million in goods price plus overdue interest"
Administrative penalty decisionDecides to impose..."Decides to order the party to correct the unlawful conduct and imposes a fine of RMB 500,000"

Key principle: Accurately distinguish "parties' claims" from "the instrument's conclusions." Do not misstate the plaintiff's claims as the court's judgment.

Step 6: Extract Legal Authorities

List the legal norms cited in the instrument:

  • Statutes: Civil Code Art. X; Criminal Law Art. X
  • Judicial interpretations: SPC Interpretation on ... Art. X
  • Administrative regulations / departmental rules
  • Contract clauses (if a contract dispute)
  • Arbitration rules (arbitral awards)
  • Administrative discretion benchmarks (administrative penalty decisions)

Notes:

  • Prioritize norms directly cited and used as the basis for judgment / award / penalty
  • If a cited provision has been repealed or amended, that may be noted in the summary
  • Norms that are "referred to" or "consulted" may be listed separately
Step 7: Summarize the Reasoning

Outline in concise language the logical process by which the conclusion was reached:

The reasoning summary should show:

  1. Which facts the instrument accepted (evidence assessment)
  2. How legal norms were applied (interpretation and application of provisions)
  3. How main disputes and defenses were addressed (responses to opposing arguments)
  4. How the conclusion was derived from facts + norms (logical chain)

Requirements:

  • Not a mere abridgment of the original, but a demonstration of "how facts were found → how norms were applied → how the conclusion was reached"
  • Concise language highlighting the causal chain
  • Avoid restating the full pleadings history

Phase Three: Output

Step 8: Format Output by Document Type

According to the identified document type, select the corresponding structured template for output (see output format templates in Part III).

III. Output Format Templates

Template A: General Six-Element Summary (Applicable to Any Legal Instrument)

markdown
## Case Summary

### I. Case Information
- **Case name:** [name]
- **Case number:** [number]
- **Date:** [date]
- **Issuing body:** [body]
- **Parties:** [plaintiff / claimant / sanctioned party] vs [defendant / respondent / sanctioning authority]

### II. Case Facts (or Procedural Background)
[Concise statement of fact points directly related to the conclusion; use third person; retain only key information]

### III. Issues in Dispute
1. [Issue 1, phrased as a legal question]
2. [Issue 2, if any]

### IV. Conclusion
- [Conclusion on Issue 1 is...]
- [Conclusion on Issue 2 is...]

### V. Legal Authorities
- [XXX Law] Art. X
- [judicial interpretation / regulation / contract clause / arbitration rule]

### VI. Reasoning
[Outline the logical path from fact findings through application of law to the conclusion; highlight the causal chain]

Template B: Judgment Summary

markdown
## Judgment Summary

### I. Case Information
- **Case name:** [...]
- **Case number:** [...]
- **Court:** [...]
- **Decision date:** [...]

### II. Case Facts
[Cause of action, parties' identities and main claims; retain only key facts affecting the judgment]

### III. Issues in Dispute
1. [...]

### IV. Facts Found by the Court
[The court's findings on key facts (as distinct from parties' allegations)]

### V. Legal Authorities
- [...]

### VI. Disposition
- [Specific content of grant / dismissal / confirmation / modification]

### VII. Reasoning
[The court's logical chain from evidence → fact findings → application of law → conclusion]

Template C: Ruling Summary

markdown
## Ruling Summary

### I. Case Information
- **Case name:** [...]
- **Case number:** [...]
- **Court:** [...]
- **Ruling date:** [...]

### II. Procedural Background
[Procedural stage of the case, e.g., after the plaintiff sued, the defendant raised a jurisdictional objection]

### III. Content of the Application or Objection
[Specific requests or assertions of the applicant / objector]

### IV. Issues in Dispute
[Procedural question to be resolved]

### V. Legal Authorities
- [...]

### VI. Ruling Result
- [Grant / dismiss / transfer / stay / terminate]

### VII. Reasoning
[The court's analytical logic on the procedural question]

Template D: Mediation Statement Summary

markdown
## Mediation Statement Summary

### I. Case Information
- **Case name:** [...]
- **Case number:** [...]
- **Court:** [...]
- **Mediation date:** [...]

### II. Case Background
[Matters in dispute and the background of reaching mediation]

### III. Mediation Matters
[Core matters in dispute between the parties]

### IV. Content of the Mediation Agreement
- Article 1: [...]
- Article 2: [...]
- [Manner of performance, deadlines, liability for breach, etc.]

### V. Legal Authorities
- [Legal basis confirming the validity of the mediation agreement]

### VI. Result
[The mediation agreement is confirmed by the court and has legal effect]

Template E: Arbitral Award Summary

markdown
## Arbitral Award Summary

### I. Case Information
- **Case name:** [...]
- **Case number:** [...]
- **Arbitration institution:** [...]
- **Award date:** [...]
- **Tribunal composition:** [sole arbitrator / three-member tribunal]

### II. Case Facts
[Arbitration claims and disputed facts; retain only content related to the award]

### III. Issues in Dispute
[Core questions the tribunal needed to resolve]

### IV. Facts Found by the Tribunal
[The tribunal's findings of fact]

### V. Legal Authorities
- [arbitration agreement / arbitration rules]
- [substantive legal norms]

### VI. Award Disposition
- [...]
- [Note on finality of one award]

### VII. Reasoning
[The tribunal's analytical logic]

Template F: Administrative Penalty Decision Summary

markdown
## Administrative Penalty Decision Summary

### I. Case Information
- **Penalty decision document number:** [...]
- **Issuing authority:** [...]
- **Date:** [...]

### II. Party Information
- **Sanctioned party:** [name]
- **Unified social credit code / ID number:** [...]

### III. Unlawful Facts
[Unlawful facts found upon investigation; state objectively]

### IV. Issues in Dispute (if statements or defenses were made)
[The party's defense and the agency's response]

### V. Legal Authorities
- [Basis for finding unlawfulness]
- [Basis for the penalty]
- [Discretion benchmarks for the penalty]

### VI. Penalty Result
- [Type of penalty: fine / order to correct / revoke license, etc.]
- [Measure / amount]
- [Manner and deadline for performance]

### VII. Reasoning
[The agency's explanation of fact findings, application of law, and discretion as to the penalty measure]

### VIII. Remedies
- **Administrative reconsideration:** [deadline, authority]
- **Administrative litigation:** [deadline, court with jurisdiction]

IV. Examples

Example 1: Civil Judgment Summary

Input (key points from the original):

Beijing Haidian District People's Court Civil Judgment (2024) Jing 0108 Min Chu No. 1234. Plaintiff Meijing Co. sued defendant Ant Co. for unfair competition. Plaintiff alleged that Ant Co. scraped data from plaintiff's "XX Data Platform" by technical means for use in its own product operations, constituting unfair competition. Defendant argued that it scraped publicly available data and had taken reasonable measures, and thus did not constitute unfair competition. Court findings: Defendant scraped plaintiff's platform data without permission; although some of the data was public, substantial cost was required to collect and organize it; defendant's conduct substantially substituted for plaintiff's data product, harmed plaintiff's lawful rights and interests, and disrupted market competition order. Judgment: (1) Defendant shall immediately cease the unfair competition at issue; (2) Defendant shall compensate plaintiff RMB 2 million for economic losses and RMB 100,000 for reasonable expenses.

Output:

markdown
## Judgment Summary

### I. Case Information
- **Case name:** Meijing Co. v. Ant Co. unfair competition dispute
- **Case number:** (2024) Jing 0108 Min Chu No. 1234
- **Court:** Beijing Haidian District People's Court
- **Decision date:** [date in original]

### II. Case Facts
Plaintiff Meijing Co. operates the "XX Data Platform" and invested substantial cost to collect and organize data. Defendant Ant Co. scraped data from plaintiff's platform by technical means for use in its own product operations. Plaintiff sued, alleging that defendant's conduct constituted unfair competition.

### III. Issues in Dispute
1. Whether defendant's scraping of plaintiff's platform data constitutes unfair competition?
2. If it constitutes unfair competition, how should the amount of damages be determined?

### IV. Facts Found by the Court
- Defendant scraped plaintiff's platform data without permission
- Although some of the data at issue was public, substantial cost was required to collect and organize it
- Defendant's conduct substantially substituted for plaintiff's data product
- Defendant's conduct harmed plaintiff's lawful rights and interests and disrupted market competition order

### V. Legal Authorities
- Anti-Unfair Competition Law Art. 2 (general clause)
- Anti-Unfair Competition Law Art. 12 (internet-specific clause)

### VI. Disposition
1. Defendant shall immediately cease the unfair competition at issue
2. Defendant shall compensate plaintiff RMB 2 million for economic losses
3. Defendant shall compensate plaintiff RMB 100,000 for reasonable expenses

### VII. Reasoning
The court found that plaintiff enjoyed competitive interests in its platform data; defendant's scraping without permission and substantial substitution for plaintiff's data product violated the principle of good faith and recognized commercial ethics, disrupted market competition order, and constituted unfair competition. As to the amount of damages, the court comprehensively considered the market value of plaintiff's data product, the nature and duration of defendant's infringement, and other factors, and assessed damages at RMB 2 million.

Example 2: Ruling Summary

Input (key points from the original):

Beijing No. 1 Intermediate People's Court Civil Ruling. Plaintiff Meijing Co. sued defendant Ant Co. for unfair competition in Haidian District Court. Within the answer period, defendant raised a jurisdictional objection, asserting that the case should be heard by the court at defendant's domicile (Chaoyang District, Beijing). Haidian District Court held that this was an unfair competition dispute and that the court at the place of infringement (Haidian District) had jurisdiction, and ruled to dismiss the defendant's jurisdictional objection. Defendant appealed. The Intermediate Court held that the alleged infringement occurred in Haidian District, and Haidian District Court, as the court at the place of infringement, had jurisdiction. Ruling: appeal dismissed; original ruling affirmed.

Output:

markdown
## Ruling Summary

### I. Case Information
- **Case name:** Meijing Co. v. Ant Co. unfair competition dispute (appeal on jurisdictional objection)
- **Case number:** [case number]
- **Court:** Beijing No. 1 Intermediate People's Court
- **Ruling date:** [date]

### II. Procedural Background
Plaintiff sued in Haidian District Court; defendant raised a jurisdictional objection; the court of first instance ruled to dismiss. Defendant appealed.

### III. Content of the Application or Objection
Appellant (defendant at first instance) asserted that the case should be heard by the court at defendant's domicile (Chaoyang District Court).

### IV. Issues in Dispute
Should the case be heard by the court at the defendant's domicile, or by the court at the place of infringement?

### V. Legal Authorities
- Civil Procedure Law Art. 29 (jurisdiction at the place of infringement)
- Civil Procedure Law Art. 130 (review of jurisdictional objections)

### VI. Ruling Result
Ruling dismissing the appeal and affirming the original ruling (Haidian District Court has jurisdiction over this case).

### VII. Reasoning
The appellate court held that this was an unfair competition dispute and that the alleged infringement occurred in Haidian District. Under Civil Procedure Law Art. 29, actions arising from infringement are heard by the court at the place of infringement or at the defendant's domicile. Haidian District, as the place of infringement, has jurisdiction. The first-instance ruling was correct and should be affirmed.

Example 3: Administrative Penalty Decision Summary

Input (key points from the original):

Beijing Municipal Administration for Market Regulation Administrative Penalty Decision. Party: a certain technology company Ltd. Upon investigation, the party collected users' clipboard information in its operated App without user consent, for identifying user acquisition channels. This violated Personal Information Protection Law Arts. 13 and 16. The party defended that it collected information to optimize user experience and had applied anonymization. The agency held that clipboard information is personal information; collection without prior user consent violated the minimum-necessity principle; the anonymization defense failed. Penalty decision: order correction of the unlawful conduct and impose a fine of RMB 500,000.

Output:

markdown
## Administrative Penalty Decision Summary

### I. Case Information
- **Penalty decision document number:** [number]
- **Issuing authority:** Beijing Municipal Administration for Market Regulation
- **Date:** [date]

### II. Party Information
- **Sanctioned party:** a certain technology company Ltd.
- **Unified social credit code:** [...]

### III. Unlawful Facts
In its operated App, the party collected users' clipboard information without user consent for identifying user acquisition channels.

### IV. Issues in Dispute
1. Is clipboard information personal information subject to protection?
2. Do the party's defenses based on "optimizing user experience" and "anonymization" succeed?

### V. Legal Authorities
- **Basis for finding unlawfulness:** Personal Information Protection Law Art. 13 (processing personal information requires the individual's consent); Art. 16 (may not refuse to provide products or services on the ground that the individual does not consent to processing or withdraws consent)
- **Basis for penalty:** Personal Information Protection Law Art. 66

### VI. Penalty Result
- Order correction of the unlawful conduct
- Fine of RMB 500,000

### VII. Reasoning
The agency found: clipboard information can identify a specific natural person and constitutes personal information. The party collected it without obtaining user consent beforehand, and the collection exceeded what was necessary to achieve product functions, violating the minimum-necessity principle. The purpose of "optimizing user experience" does not exempt the consent obligation, and the "anonymization" defense was inconsistent with the facts. Comprehensively considering the nature, circumstances, and degree of harm of the unlawful conduct, a fine of RMB 500,000 was imposed.

### VIII. Remedies
- **Administrative reconsideration:** Apply to [reconsideration authority] within 60 days of receipt of the decision
- **Administrative litigation:** File with [court of jurisdiction] within 6 months of receipt of the decision

V. Common Errors and Prevention

Error TypeError DescriptionConsequencesPreventive Measures
Wrong document-type identificationSummarizing a ruling in judgment format and extracting lengthy substantive disputesSummary emphases completely off; true procedural issues missedIdentify document type first, then select the matching template
Confusing allegations with findingsWriting facts alleged by the plaintiff as "facts found by the court"Inaccurate summary; misleads readersStrictly distinguish "parties' allegations" from "instrument findings"; mark the source of the latter
Overly broad issues"The parties had a dispute" / "there is a controversy"Fails to reveal the legal questions the case truly needed to resolvePhrase issues as answerable legal questions
Mixing in subjective evaluation"The court wisely held" / "the penalty was quite reasonable"Violates objectivity and neutralityDelete all evaluative adjectives; state only facts and conclusions
Restating rather than summarizingLong quotations from the original with only light cutsLoses the value of a summary; remains lengthyExtract core elements, then reorganize in summarizing language
Omitting legal authoritiesWriting only "pursuant to relevant legal provisions"Cannot trace the normative basis of the decision / penaltyList specific provisions cited, item by item
Empty reasoning"The court rendered judgment according to law"No logical chain; readers cannot see how the conclusion was reachedShow the full chain "fact findings → application of norms → derivation of conclusion"
Ignoring procedural boilerplateRetaining lengthy service history, hearing notices, and collegial panel compositionInflates length; core information is buriedCompress or omit procedural content that does not affect the conclusion

VI. Special Scenario Handling

6.1 Multi-Defendant / Multi-Plaintiff Cases

  • Clearly list all parties' litigation status in the "Parties" section
  • Distinguish each party's different claims and conduct in "Case Facts"
  • Separately list dispositions as to each party in "Disposition"

6.2 Series / Related Cases

  • At the start of the summary, mark the relationship between this case and related cases
  • If the disposition depends on another case's judgment, state that
  • For batch litigation (e.g., securities misrepresentation series), simplify common facts and highlight case-specific differences

6.3 Second-Instance / Retrial Instruments

  • Second-instance judgment summary: Focus on differences between first-instance findings and second-instance modification / affirmance, and the appellate court's reasons for modification
  • Retrial instrument summary: Focus on the parties' grounds for applying for retrial and the court's review conclusions regarding the prior proceedings
  • Explain the instance history in "Procedural Background"

6.4 Scanned / OCR-Recognized Instruments

  • Where recognition of text is uncertain, mark with [?] in the summary
  • Especially mark critical numbers (amounts, dates, case numbers) if unclear
  • Advise the user to verify against the original, especially article numbers of legal authorities

6.5 Foreign-Related / Hong Kong–Macao–Taiwan Cases

  • Note the applicable law (Chinese law, foreign law, international treaties)
  • If conflict-of-laws rules are involved, state the governing law chosen by the court
  • For rulings involving service abroad, taking of evidence abroad, and similar procedural issues, extract these as emphases

6.6 Class Actions / Representative Actions

  • In "Parties," explain the litigation representatives and the represented plaintiff group
  • In "Case Facts," summarize the common factual basis
  • In "Disposition," state the scope of effect of the judgment

VII. Quality Checklist

markdown
□ 1. Has the document type been accurately identified (judgment / ruling / mediation / arbitration / administrative penalty)?
□ 2. Is case information complete (name, case number, date, issuing body, parties)?
□ 3. Does the facts section retain only information directly related to the conclusion?
□ 4. Have "parties' allegations" and "instrument findings" been strictly distinguished?
□ 5. Have issues in dispute been distilled into clear legal questions?
□ 6. Does the conclusion accurately reflect the instrument's final disposition?
□ 7. Have legal authorities been listed with specific provisions (statute name + article number)?
□ 8. Does the reasoning show the logical chain "facts → norms → conclusion"?
□ 9. Is the summary free of any subjective evaluative language?
□ 10. Does the output format match the document type (correct template used)?
□ 11. Has procedural boilerplate been appropriately compressed?
□ 12. Have uncertainties from scanned / OCR recognition been marked?

VIII. Related Skills

Related SkillRelationshipNotes
Legal document formattingComplementaryEnsure normative summary format and clean layout
Statutory retrievalUpstream / parallelVerify that legal authorities extracted in the summary remain currently in force
Legal reasoningDownstreamFurther analyze reasoning logic and argumentative structure on the basis of the summary
Legal concept comprehensionUpstreamUnderstand key legal concepts appearing in the summary (e.g., "public order and good morals," "good faith")
Similar-case retrieval and analysisDownstreamCompare and analyze after summarizing multiple instruments
Multi-document summarizationExtensionWhen batch summarization and comparison of multiple instruments are needed

IX. Limitations and Risk Notices

  • Dependence on information completeness: Summary quality depends entirely on the completeness and clarity of the original. If the original is logically confused or insufficiently reasoned, the summary cannot cure that.
  • OCR recognition risk: OCR of scans may misrecognize characters, numbers, or punctuation; for critical information, advise the user to verify against the original.
  • Subjective understanding bias: Although summaries require objectivity and neutrality, selecting "key facts" and distilling "issues in dispute" still inevitably involve judgment; different persons may emphasize different points.
  • Format limits: Standardized templates cannot cover all special document types; for novel or uncommon instruments, flexibly adjust on the basis of the general six-element framework.
  • Currency: The currency of legal authorities is judged as of the instrument's issuance; provisions cited in the summary may later have been amended; if used for current legal analysis, verify separately.
  • Confidential information: For cases involving closed hearings, trade secrets, or personal privacy, the summary must observe disclosure limits and, where necessary, redact sensitive information.

Frequently asked questions

What does the Legal Document Summarization AI skill do?

Trigger this skill when a structured summary is needed for judgments, rulings, mediation statements, arbitral awards, administrative penalty decisions, and other legal instruments. Typical trigger scenarios include, but are not limited to: - The user asks for a summary of a legal instrument, extraction of key points, or generation of a holding / ratio abstract - The user needs to convert multiple legal instruments into standardized summaries that are comparable and searchable - The user asks to extract case information, issues in dispute, legal authorities, or reasoning - The user needs sum...

Why use Legal Document Summarization on TypingMind?

Because you install it once and use it with any model. Legal Document Summarization is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Legal Document Summarization in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/THUYRan/Legal-Skills-Chinese/tree/main/English/skills/legal-document-summarization. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Legal Document Summarization?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Legal Document Summarization?

As many as you like. As long as a model supports skills, you can use Legal Document Summarization with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Legal Document Summarization AI skill free?

It is published on GitHub by THUYRan. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇