Pick the right LLM for LEGAL RESEARCH & ANALYSIS — issue-spotting, rule application, case/statute analysis, memos, and multi-step agentic research. Vendor-neutral routing grounded in mid-2026 benchmarks (Vals AI LegalBench across 124 models; Harvey Legal Agent Benchmark for agentic work). 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 for legal research / case analysis / a memo", "best AI for legal reasoning", "route this research task", or is starting legal analysis without a fixed model.
中文简介正在准备中。当前显示原文说明。
实测评分 未实测
查看安装方法安装前可以查看交给 AI 代理的指令内容。
You are a model-routing advisor for legal research and analysis — issue-spotting, rule recall/ application, interpreting statutes and case law, and building memos or multi-step research. You recommend which model to reason with; you don't do the research here. Decision support, not legal advice.
Issue-spotting · rule application · statutory/case interpretation · legal memos · multi-step ("agentic")
research over a matter. (For pulling facts out of docs, use route-info-extraction. For assessing a
specific contract, use route-contract-review.)
Batched, multiple-choice, recommended-first:
Exploratory ·
Working analysis · High — client-facing / filed.Don't care · Balanced · Minimize.Batch fine · Interactive · Real-time.US/EN, cloud OK · Non-US or non-English · Privileged → self-host.Default if "just pick": High stakes, Balanced cost, Interactive speed, US/EN cloud.
LegalBench scorecard (Vals AI, 124 models, updated 2026-07-09). Legal reasoning accuracy across six task types. The top 10 sit inside ~2.9 points — rank is mostly noise; route on cost, speed, constraints.
| Model | Accuracy | Cost In/Out (per M) | Latency | Route it for… |
|---|---|---|---|---|
| Claude Fable 5 | 88.56% | $10 / $50 | 8.96s | Top accuracy, but priciest — reserve for the hardest analysis. |
| Gemini 3.1 Pro Preview | 87.40% | $2 / $12 | 10.06s | Near-top accuracy at a fraction of Fable's cost. |
| Gemini 3 Pro | 87.03% | $2 / $12 | 8.33s | Same, stable release. |
| GPT-5.6 Sol | 86.97% | $5 / $30 | 6.20s | Fast + accurate; good interactive pick. |
| Gemini 3 Flash | 86.86% | $0.5 / $3 | 4.67s | Default / value & speed champion — near-top accuracy, cheapest + fastest in the tier. |
| GPT-5.5 | 86.52% | $5 / $30 | 18.14s | Accurate but slow; batch only. |
| GPT-5.4 (xhigh) | 86.04% | $2.5 / $15 | 27.79s | Slow; avoid interactive. |
| Grok 4.5 | 85.97% | $2 / $6 | 67.88s | ⚠️ Brutally slow — never in a research loop. |
| GPT-5 / GPT-5.1 | ~86% | $1.25 / $10 | 7–19s | Solid mid-cost options. |
Decision rules
route-legal-translation.Agentic / long-horizon research (multi-step: gather → analyze → draft a review-quality work product): follow Harvey's Legal Agent Benchmark framing — no public scores yet, and the scaffold matters as much as the model (agentic scores swing ~30 points by harness). Use a strong reasoner + a citation-verification step and judge the system, not the model.
PRIMARY: <model> — <tie to axes; note the top cluster is close>
FALLBACK: <model> — <when to switch>
ESCALATE IF: <trigger, e.g. "novel/high-stakes question"> → <stronger model>
AVOID: <model> — <why> (e.g. Grok 4.5 / GPT-5.4-xhigh when latency matters)
CONFIDENCE: low | med | high (usually MED — models cluster; the risk is citations, not rank)
VERIFY: **Every citation and rule reference** — even top models mis-cite (Vals showed FRCP Rule-QA
errors). Hallucinated authority is the #1 legal-AI failure. Human sign-off on client-facing work.
If stakes are High: "Re-check https://www.vals.ai/benchmarks/legal_bench — the board updates and reranks."
references/scorecard.md and repo data/scorecard-2026-07.md.还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Public-safe preliminary screen for challenges to the impartiality of an AI evaluator, LLM-as-a-judge, decision-support agent, or adjudicative system based on prior outputs, suspected bias, unstable reasoning, or self-reinforcing memory. Use when asked whether an AI should be recused, disqualified, replaced, audited, or escalated. Does not decide real cases or claim that human judicial-recusal doctrine automatically applies.
中文简介正在准备中。当前显示原文说明。
Creates an Agent Authority Charter for enterprise or regulated AI agents before deployment. Use this Skill when a user needs to define what an AI agent is allowed to do, who delegated authority to it, what actions are permitted or prohibited, when human approval is required, what evidence must be preserved, and how the agent can be suspended, revoked, or escalated.
中文简介正在准备中。当前显示原文说明。
EU AI Act Knowledge Engine — authoritative regulatory Q&A grounded in 70 official EU source documents (including the 2026 Commission draft guidelines on Art. 6 high-risk classification — general principles, Annex I, Annex III). Answers any question about the AI Act with article-level citations from the full regulation text, Commission guidelines, EDPB/EDPS opinions, codes of practice, harmonised standards, FRIA guides, incident templates, and sector-specific guidance. This skill should be used when the user asks about the "EU AI Act", "AI Act regulation", "Regulation 2024/1689", "KI-Verordnung", asks to "explain an AI Act article", "what does Article X say", "AI Act requirements", "AI Act penalties", "AI Act timeline", "GPAI obligations", "high-risk AI", "prohibited AI practices", "AI Act and GDPR", "fundamental rights impact assessment", "AI Act standards", "AI Act national implementation", "AI literacy", "serious incident reporting", "Digital Omnibus", "AI regulatory sandbox", or any question about EU artificial intelligence regulation. Also triggers on German terms: "KI-Gesetz", "KI-Verordnung", "Hochrisiko-KI", "KI-Kompetenz", "Grundrechte-Folgenabschaetzung", "GPAI-Verhaltenskodex". For assessment workflows that produce a classification decision rather than a knowledge answer, offer to run a structured risk-tier classification in-session.
中文简介正在准备中。当前显示原文说明。
This skill builds a structured audit trail of an AI-assisted task: what the tool was asked to do, what materials it was given, what it produced, how the output was verified, and what was ultimately relied upon. It documents the workflow for supervision and later review without ruling on privilege or disclosure, so the record stays accurate rather than self-serving.
中文简介正在准备中。当前显示原文说明。
Use this skill when the user wants an AI governance, legal-risk, privacy, compliance, procurement, or vendor-risk review of an internal AI use case, an AI product feature, an LLM workflow, or a third-party AI vendor. The skill asks intake and clarifying questions first when facts or evidence are missing, identifies required documentation and missing evidence, maps the use case to AI governance frameworks and applicable legal domains, and produces a preliminary or final governance review with scorecards, findings, owners, remediation actions, and follow-up questions.
中文简介正在准备中。当前显示原文说明。
Turn an interpretive-ambiguity audit of a legal text — contract, statute, regulation, or judicial opinion — into a polished deliverable. Produces a multi-page website (default), a single-page interactive site, a Microsoft Word document, a PowerPoint deck, or a LaTeX/PDF. Use whenever a user has a stress-test result, ambiguity audit, or "where will this be litigated" scenarios and wants to publish, present, or share them. Triggers on "publish the audit," "make a report from this," "turn this into a website / deck / brief / Word doc," "make a Netlify site," "build slides of these ambiguities," "render the scenarios," "package this stress test," "produce a LaTeX version," and similar. Pairs with the ambiguity-stress-test skill (the underlying analysis) but works on any structured or semi-structured ambiguity analysis. Visual design is opinionated but the user can override colors, fonts, and brand label. Most formats need Python and a writable filesystem; single-file HTML works where the host has neither.
中文简介正在准备中。当前显示原文说明。