Vendor AI Review

Review AI vendor contracts (LLM API providers, ML libraries, AI services). Включает SaaS-MSA общие issues (см. contract-law/saas-msa-review) + AI-specific: training data rights, output ownership, model warranty, hallucination liability, rate limits, API stability, безопасность (запрет training на our data), exit strategy + portability. Post-2022 sanctions specifics.

43 stars 7 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add AlsKozlov/ru-legal --skill vendor-ai-review
Manual install
git clone --depth 1 https://github.com/AlsKozlov/ru-legal.git /tmp/ru-legal
mkdir -p ~/.claude/skills
cp -r /tmp/ru-legal/packs/ai-governance/skills/vendor-ai-review ~/.claude/skills/vendor-ai-review

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# /vendor-ai-review

## Назначение

Review AI vendor / API / library contracts. Differs from generic SaaS-MSA review
(см. `contract-law/saas-msa-review`) — focus на AI-specific clauses.

## Pre-flight

- **Vendor identified:** [имя + type: hyperscaler LLM API / Russian LLM / open-source / SDK / library]
- **Sanctions exposure check:** US/EU vendor with restricted РФ access? (post-2022)
- **Use case clear:** [для каких задач будем использовать AI]
- **Data sensitivity:** [какие данные будем sending к vendor]

## AI-specific contract review points

### A. Data rights — наши данные у vendor

#### Training on our data

**Critical:** Vendor может использовать наш input / output для training future models?

- **Best:** "Vendor shall not use Customer Data for training models" — explicit
- **OK:** Opt-out available + opt-out exercised
- **Problematic:** Default training on customer data with no opt-out
- **Bad:** Training explicit + irrevocable license к vendor

**Examples in industry:**
- **OpenAI Enterprise:** Не trains on data (default for enterprise)
- **OpenAI consumer ChatGPT:** Trains by default (opt-out exists)
- **Anthropic Claude (API):** Не trains by default
- **Google Bard / Gemini Enterprise:** Не trains by default
- **YandexGPT API:** Check current terms (varies)
- **GigaChat:** Check current terms

→ Push for explicit "no training" в SaaS-MSA.

#### Data retention

- Как долго vendor stores наши queries / responses?
- Recommended: short retention (30 days max)
Read the full skill on GitHub

Details

Skill name
vendor-ai-review
Path
packs/ai-governance/skills/vendor-ai-review/SKILL.md
Repo updated
26 May 2026

FAQ

What does the Vendor AI Review skill do?

Review AI vendor contracts (LLM API providers, ML libraries, AI services). Включает SaaS-MSA общие issues (см. contract-law/saas-msa-review) + AI-specific: training data rights, output ownership, model warranty, hallucination liability, rate limits, API stability, безопасность (запрет training на our data), exit strategy + portability. Post-2022 sanctions specifics.

How do I install the Vendor AI Review skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add AlsKozlov/ru-legal --skill vendor-ai-review. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Vendor AI Review popular?

It lives in AlsKozlov/ru-legal, which has 43 GitHub stars and 7 forks.

Related skills

Administrative Appeal

43

AlsKozlov/ru-legal

Обжалование постановления о привлечении к административной ответственности. КоАП ст.30.3 — **10 суток** с момента вручения постановления. Two-tier: обжалование к вышестоящему должн.лицу / органу ИЛИ в суд; затем доп.tier. Critical: 10-day clock — не пропустить!

Other

AI Inventory

43

AlsKozlov/ru-legal

Inventory всех AI systems в организации — где, какие, для чего, на каких данных, какой compliance status. Critical для regulatory compliance + risk management.

AI, LLMs & Agents

Aia Generation

43

AlsKozlov/ru-legal

AI Impact Assessment (AIA) — formal documentation для AI use case рассматриваемого как medium/high risk. РФ specifics: AIA combines AI-specific risks с ОВПД framework (152-ФЗ ст.18.1) если processing ПДн. Структура: scope + risks + mitigation + approval. Subject to compliance officer / AI Officer / outside-адв ФПА review.

AI, LLMs & Agents

Verify Agent Output

45

pjt222/agent-almanac

Validate deliverables and build evidence trails when work passes between agents. Covers expected outcome specification before execution, structured evidence generation during execution, deliverable validation against external anchors after execution, fidelity checks for compressed or summarized outputs, trust boundary classification, and structured disagreement reporting on verification failure. Use when coordinating multi-agent workflows, reviewing cross-agent handoffs, producing external-facing outputs, or auditing whether an agent's summary faithfully represents its source material.

AI, LLMs & AgentsIn 1 list

Use Case Triage

43

AlsKozlov/ru-legal

Triage AI use case по уровню риска — PROCEED / IMPACT ASSESS / RESTRICT / PROHIBIT. РФ specifics: combination 152-ФЗ ПДн + sectoral overlay (Банк России для финансовых; Минздрав для мед; Минобрнауки для education; и т.д.) + национальная стратегия AI (Указ Президента 490/2019).

AI, LLMs & Agents

Version Ml Data

45

pjt222/agent-almanac

Version machine learning datasets using DVC (Data Version Control) with remote storage backends, build reproducible data pipelines with dependency tracking, integrate with Git workflows, and ensure data lineage for model reproducibility. Use when versioning large datasets that do not fit in Git, tracking data changes alongside code changes, ensuring ML experiment reproducibility, sharing datasets across team members, or auditing data lineage for compliance requirements.

AI, LLMs & AgentsIn 1 list

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.