Data & AI skills

Design a learning sequence teaching prompt quality — comparing vague vs. refined prompts to show why specificity and context transform AI output. Use when students use AI without understanding why output quality varies.

The English summary is being prepared. Showing the original description.

Create self-explanation prompts that deepen understanding of worked examples, texts, or diagrams. Use when students read material passively without engaging with underlying principles.

The English summary is being prepared. Showing the original description.

Interpret learning analytics data and translate dashboard findings into actionable teaching decisions. Use when reviewing LMS data, quiz patterns, or engagement metrics.

The English summary is being prepared. Showing the original description.

Use when the user has a RAG (Retrieval-Augmented Generation) system and wants to evaluate its quality — separating retrieval issues from generation issues. Also use when the user mentions RAG evaluation, faithfulness checking, hallucination detection in RAG, retrieval quality, chunking optimization, or "is my RAG pipeline working." Outputs a diagnostic matrix that pinpoints whether problems are in retrieval or generation.

The English summary is being prepared. Showing the original description.

Use when the user has a judge/grader and human-labeled data, and wants to measure how well the judge agrees with humans, detect systematic biases, determine whether automatic evaluation can replace human review, or build a human-reduction roadmap. Also use when the user mentions calibration, TPR/TNR, judge validation, inter-rater agreement, Cohen's kappa, bias detection, or "is my automatic evaluation trustworthy." Merges the calibrate and align functions into one skill.

The English summary is being prepared. Showing the original description.

Benchmark LLM reference recommendation capabilities by verifying every cited paper against Crossref, PubMed, arXiv, and DBLP. Measures hallucination rate, per-field accuracy (title/author/year/DOI), discipline breakdown, and year constraint compliance. Supports tool-augmented (ReAct + web search) mode. Use when the user asks to evaluate, benchmark, or compare models on academic reference hallucination, literature recommendation quality, or citation accuracy.

The English summary is being prepared. Showing the original description.

Use when the user has changed a prompt (system prompt, RAG template, agent instruction, etc.) and wants to know whether the candidate is better or worse than the baseline. Also use when the user mentions prompt A/B testing, prompt comparison, prompt optimization validation, "did my prompt change help," or prompt regression testing. Outputs per-dimension win rates with statistical significance using OpenJudge PairwiseAnalyzer.

The English summary is being prepared. Showing the original description.

Build RL reward signals using the OpenJudge framework. Covers choosing between pointwise and pairwise reward strategies based on RL algorithm, task type, and cost; aggregating multi-dimensional pointwise scores into a scalar reward; pairwise tournament reward for GRPO on subjective tasks (net win rate across group rollouts); generating preference pairs for DPO/RLAIF; and normalizing scores for training stability. Use when building reward models, scoring rollouts for GRPO/REINFORCE, generating preference data for DPO, or doing Best-of-N selection.

The English summary is being prepared. Showing the original description.

Build custom LLM evaluation pipelines using the OpenJudge framework. Covers selecting and configuring graders (LLM-based, function-based, agentic), running batch evaluations with GradingRunner, combining scores with aggregators, applying evaluation strategies (voting, average), auto-generating graders from data, and analyzing results (pairwise win rates, statistics, validation metrics). Use when the user wants to evaluate LLM outputs, compare multiple models, design scoring criteria, or build an automated evaluation system.

The English summary is being prepared. Showing the original description.

Drafts a customized Customer Agreement starting from the Y Combinator standard form SaaS template. Tailors the agreement through structured intake questions covering fee structure, data handling, ML rights, implementation services, and more. Applies 18 always-on defaults that transform the raw YC form into a professional starting point (renamed to "Customer Agreement", data privacy section added, warranty restructured, consolidated SLA/support exhibit, etc.). Produces a clean .docx and a lawyer-facing memo explaining every change from the YC standard. Use when user says "draft a SaaS agreement", "YC SaaS", "startup SaaS contract", "customer agreement", "SaaS subscription agreement", or "I need a SaaS agreement starting from the YC form". Also trigger when the user is a startup founder discussing SaaS contracting, even if they don't mention YC specifically.

The English summary is being prepared. Showing the original description.

Pick the right LLM for LEGAL INFO EXTRACTION — pulling facts, clauses, dates, parties, obligations, and structured fields out of contracts and legal documents. Vendor-neutral routing grounded in mid-2026 benchmarks (legalbenchmarks.ai Info Extraction; CUAD/MAUD/ACORD). Asks up to 4 quick questions (cost, speed, accuracy/stakes, privacy/jurisdiction/language), then recommends a primary model + fallback + what to avoid + what a human must verify. Use when someone asks "which model should I use to extract clauses/data from these documents", "best AI for contract data extraction", "route this extraction task", or is about to pull structured fields from legal docs without a fixed model.

The English summary is being prepared. Showing the original description.

Generates standalone interactive HTML "deal cards" that translate complex regulations into negotiation-ready reference tools, systematically distinguishing mandatory obligations from negotiable implementation choices. Use when the user needs an interactive regulatory guide for (1) contract negotiation support, (2) client education or internal training, (3) regulatory briefings for commercial stakeholders, or (4) structured comparison between required and flexible compliance paths. Primary focus on EU digital regulation (Data Act, AI Act, CRA, DORA, NIS2, GDPR) but the structural pattern transfers to any regulation where separating hard obligations from implementation choice is the point. Supports bilingual output where the jurisdiction calls for it.

The English summary is being prepared. Showing the original description.

Draft matter status reports from emails, call notes, and updates. Internal and client-facing formats, RAG logic, variance commentary, escalation flags. Use when asked to draft a status report, write a project update, summarise matter progress, prepare a client report, create a weekly or monthly update, convert emails into a status summary, or produce any kind of matter reporting. Also triggers when the user pastes email threads and asks what the status is, or needs to turn internal updates into client-facing reports.

The English summary is being prepared. Showing the original description.

Panel health assessment, firm exit management, coverage gap analysis, and panel refresh brief for in-house legal ops teams. Produce an annual Panel Health Report synthesising scorecard data, billing compliance, and step-out patterns into per-firm recommendations — Retain, Watch, Improvement Plan, or Exit Review. Produce a firm exit pack: formal exit notice with active matter transition plan, plus an internal exit record. Identify practice area, geographic, or tier gaps and produce a structured Coverage Gap Report with remediation options. Produce a Panel Refresh Brief as a scoped input to the RFP process. Trigger on: 'review our panel', 'annual panel review', 'is our panel working', 'which firms should we exit', 'write the exit notice', 'coverage gap', 'we don't have anyone for restructuring', 'refresh the panel', 'panel rationalisation', 'too many firms', 'step-out problem', 'do we have the right firms'.

The English summary is being prepared. Showing the original description.

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.

The English summary is being prepared. Showing the original description.

Performance scorecard design, post-matter feedback collection, QBR preparation, and firm comparison for in-house legal ops teams evaluating outside counsel. Design a scorecard framework calibrated to your team's maturity level. Produce post-matter feedback forms for supervising attorneys to complete at matter close. Prepare a QBR pack with agenda, data summary, and talking points for structured business reviews. Produce a side-by-side firm comparison for panel decision-making. Trigger on: 'design a performance scorecard', 'how do we evaluate our firms', 'build us a scorecard', 'post-matter feedback form', 'feedback at matter close', 'collect feedback on the firm', 'QBR', 'quarterly business review', 'prep for the firm review', 'what should I raise with the partner', 'compare the firms', 'which firm is performing better', 'rank our panel', 'firm performance data', 'scorecard framework', 'evaluate outside counsel', 'annual firm review', 'how are our firms performing'.

The English summary is being prepared. Showing the original description.

Matter scoping across the full pre-execution arc — organise client data into a structured brief, capture the agreed baseline, or reconstruct scope mid-flight. Use when making sense of client information before a proposal, scoping a new matter, running a kickoff, defining scope, mapping stakeholders, or inheriting a matter mid-flight. Trigger on: 'make sense of this', 'structure this for the proposal', 'scope this matter', 'new matter', 'kickoff', 'what are we doing', 'who are the stakeholders', 'what does success look like', 'matter setup', 'intake', 'I've inherited this matter', 'organise this client data'.

The English summary is being prepared. Showing the original description.

Navigate privacy regulations (GDPR, CCPA), review DPAs, and handle data subject requests. Use when reviewing data processing agreements, responding to data subject access or deletion requests, assessing cross-border data transfer requirements, or evaluating privacy compliance.

The English summary is being prepared. Showing the original description.

Reviews outside counsel invoices and related billing data for an in-house legal department, including LEDES or e-billing exports, OCGs, approved rates, discounts, budgets, AFAs, and staffing rules. Starts with internal comparisons before bringing in external data. Produces invoice review findings, MBR/QBR scorecards, dispute logs, and management reports. For demonstration purposes only and not professional advice.

The English summary is being prepared. Showing the original description.

Pre-flight privilege and work-product check for legal AI prompts. Use when the user is about to send legal content to a third-party AI surface (ChatGPT, Claude.ai, Copilot, Gemini, etc.) and wants to know whether the prompt risks blowing attorney-client privilege or work-product protection. Returns a SAFE/CAUTION/STOP band with cited factors, a discovery-impact line, and a redacted-safe rewrite.

The English summary is being prepared. Showing the original description.

Expert LGPD compliance advisor for Brazil's Lei Geral de Proteção de Dados (Law 13,709/2018). Use this skill whenever a user asks about LGPD, Brazilian data protection, ANPD, personal data processing in Brazil, data subject rights under Brazilian law, legal bases for processing, sensitive data handling, DPO appointment in Brazil, data breach notification to ANPD, LGPD penalties (fines up to 2% of revenue / R$50M), international data transfers from Brazil, Brazil-EU mutual adequacy (January 2026 — SCCs/BCRs no longer needed for Brazil-EU transfers), LGPD gap assessments, privacy policy drafting for Brazilian operations, DPIA under LGPD, consent management, or comparing LGPD with GDPR. Trigger for any Brazil privacy or data protection question even if LGPD is not named explicitly.

The English summary is being prepared. Showing the original description.

Mandatory external verification workflow for ALL non-trivial factual claims before presenting them as true. This skill MUST be used whenever Claude is asked to research any topic, answer factual questions, provide current information, draft documents containing factual claims, give legal advice or cite legal authority, discuss current events or public figures, provide technical or scientific information, state statistics or data points, or answer ANY question where the answer could have changed over time. Also trigger when the user asks Claude to 'check', 'verify', 'confirm', 'research', 'look up', or 'find out' anything, or when the user needs information that requires freshness, precision, or source accuracy. This skill overrides any default tendency to answer from memory. If a task involves stating facts, citing sources, or providing current information — use this skill.

The English summary is being prepared. Showing the original description.

Arbitrator research tool for investor-State dispute settlement (ISDS). The tool can summarize arbitrators' experience, with deliverables grounded in primary documents; it can also help find arbitrators based on the issues faced in a particular case. Use when the user asks about an arbitrator ("tell me about arbitrator X", "X, Y and Z"), wants help choosing or shortlisting arbitrators for a case, or asks about appointments, dissents, issue track records (e.g. FET, expropriation), challenge risk, double-hatting, or tribunal composition. Profiles are grounded in the UNCTAD ISDS Navigator dataset plus targeted live lookups (ICSID); deep verification retrieves the actual decisions. Sibling of isds-research.

The English summary is being prepared. Showing the original description.

Restructure any rough build, research, or legal-drafting request into an IRAC-shaped prompt — Issue, Rule, Analysis, Conclusion — optimized for a frontier model. It's the bar-exam framework, repurposed as prompt engineering. The skill leads with the issue and ends with the conclusion (where models weight attention most), forces you to name your constraints and non-goals, and specifies what "good" looks like before a single token is generated. Use it before any non-trivial build, or whenever a vague ask deserves a precise brief.

The English summary is being prepared. Showing the original description.