菜单
Use when you need to apply tracked changes and comments to DOCX files programmatically, merge multi-agent edits with conflict resolution, or produce native Word revisions from AI agent output.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
EU AI Act Quick Assessment — fast 15-25 minute triage for preliminary classification and compliance assessment. This skill should be used when the user asks to "do a quick AI Act assessment", "check if the AI Act applies to us", "run a preliminary classification", "do an AI Act triage", "quick check", "preliminary assessment", "Schnellprüfung", "Ersteinschätzung", or needs a fast initial assessment before committing to full analysis.
Use this skill when the user needs to review, draft, or redline a Data Processing Agreement (DPA / Auftragsverarbeitungsvertrag / AVV) under Art. 28 GDPR, or to prepare a Joint Controller Arrangement under Art. 26 GDPR. Triggers include "DPA", "AVV", "Auftragsverarbeitung", "Auftragsverarbeitungsvertrag", "Art. 28 contract", "data processing agreement", "processor agreement", "Art. 26 arrangement", "joint controller agreement", "JCA", "review this DPA", "draft a DPA", "redline this DPA", or any request involving controller-processor / joint-controller contracting. Supports bilingual output (DE/EN), both controller- and processor-side perspectives, and both quick (Art. 28(3)(a)–(h) coverage) and negotiation-grade (clause-by-clause risk scoring) review depths.
NIS2 Compliance Navigator — scope classification, Art. 21 gap analysis (0-4 maturity scoring), and compliance roadmap under EU Directive 2022/2555 with deep German BSIG-neu coverage and profiles for Italy, France, Netherlands, Austria, Spain. Use when: (1) User mentions "NIS2", "NIS-2", "BSIG", "BSIG-neu", "NIS2UmsuCG", "Cyberbeveiligingswet", "Loi Résilience", "NISG", "decreto legislativo 138", (2) User asks if their organization falls under NIS2 or needs a cybersecurity compliance assessment, (3) User mentions essential/important entities, Annex I/II, BSI registration, § 30 BSIG, incident reporting, management body liability, supply chain security, (4) User wants a NIS2 gap analysis, readiness assessment, or compliance roadmap, (5) User asks about NIS2 fines, enforcement, or Nachweispflicht, (6) User asks about NIS2 in any EU Member State.
Draft GDPR/DSGVO-compliant privacy notices as .docx for any EU/EEA jurisdiction and audience. Use when user asks to create a privacy policy/notice, mentions "Datenschutzerklärung", "politique de confidentialité", "privacy notice", needs Art. 13/14 disclosures, AI Act transparency, cookie policy, or notices for applicants ("Bewerber-Datenschutz"), employees ("Beschäftigten-Datenschutz"), B2B partners, or B2C customers. Covers DE (DSGVO+BDSG+TDDDG), FR (RGPD+LIL+LCEN), AT, IT, ES, NL, BE, IE, UK GDPR. Five notice types: Website/App, Applicant, Employee, Business Partner, B2C Customer.
Use whenever the user is working on document production in international arbitration: drafting requests to produce, raising or replying to objections, or preparing the schedule for the tribunal to rule on. Builds and maintains the request-to-produce table (the Redfern Schedule) for the requesting party, the producing party, or the tribunal. Applies the IBA Rules (2020) Article 3.3 admissibility checklist and the Article 9.2 objection grounds, raises a content-based political and institutional sensitivity prompt for State or state-owned parties, and writes an internal memo flagging the user's own weak requests. Trigger it even when the user does not say Redfern Schedule but mentions requests to produce, document production in an arbitration, IBA objections, Article 3.3 or 9.2, or a tribunal ruling on production. It enforces form and does not decide materiality or whether an objection will succeed.
Apply the NIST AI Risk Management Framework (NIST AI 100-1 + the NIST AI 600-1 Generative AI Profile) to a specific AI system, governance question, or impact assessment. Three modes — consult, governance plan, full assessment — all cite Subcategories (`GOVERN 1.1`) and Profile Action IDs (`GV-1.2-001`) verbatim. Use when the user mentions the AI RMF, NIST RMF, NIST AI 100-1, NIST AI 600-1, GenAI Profile, the four functions (Govern / Map / Measure / Manage), the trustworthy AI characteristics, the 12 GAI risks (confabulation, harmful bias, information integrity, CBRN, data privacy, etc.), or asks "what does NIST say about X" for an AI system.
| name | superdoc-redlines |
| description | Use when you need to apply tracked changes and comments to DOCX files programmatically, merge multi-agent edits with conflict resolution, or produce native Word revisions from AI agent output. |
| author | Yu Chou Teo |
| jurisdiction | SG |
| version | 1.0.0 |
| last_reviewed | 2026-05 |
| tags | ["docx","redlines","multi-agent","conflict-resolution","document-processing"] |
Uses stable block IDs instead of fragile text matching. Unlike approaches that try to find text by content (and fail when formatting changes), Superdoc assigns stable IDs to document blocks and applies changes by reference.
AI Agent output → Conflict resolution → Block ID mapping → Native Word revisions
import { applyRedlines } from 'superdoc-redlines';
const result = await applyRedlines({
input: './contract.docx',
edits: agentEdits,
agentName: 'Claude'
});
Audience: lawyers and developers running multi-agent DOCX workflows where several AI agents propose edits to the same document. The lawyer is the merger of last resort.
Work shape: Bounded Transactional with conflict-resolution overlay. The skill merges defined edits by stable block ID. Conflicts are surfaced, not silently resolved.
This skill provides legal support, not legal advice. The merged DOCX is an AI-assembled draft with native Word tracked changes — not a reviewed, accepted, or approved version.
Privilege and confidentiality. Runs headlessly with no required network calls. The agentName field is persisted in Word's revision XML — the AI origin of each edit is visible in the document's audit trail. Do not mask agentName as a human reviewer; that misrepresentation can affect privilege and disclosure analysis later.
Accountability. A qualified lawyer must review every tracked change in the merged output before accepting, sending, filing, or relying on it. The lawyer decides what each agent's edit is worth; the conflict-resolution engine only ensures merges are deterministic, not correct.
edits input must conform to the documented schema (per-agent, per-block).Stop and route to the responsible lawyer when: