Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
DeepSeek
Muse+22 moreClaude CodeCursorCodex CLIDeepSeekMuse
Dots
Manus
Grok
Grok Bot+18 agentsEU AI Act High-Risk Assessor
Classify AI systems against Article 6 and Annex III, then generate a legal-review-ready risk memo, high-risk obligations map, Annex IV documentation gaps, and Annex VIII registration-data draft.
$15
EU AI Act High-Risk Assessor
Example session with this skill installed
Assess our AI hiring assistant for EU AI Act high-risk status.
The local folder contains system-card.md, architecture.md, model-manifest.json, and human-review.md.
The product ranks job applicants for recruiter review based on CV data and interview responses. Recruiters see a candidate score and recommendation before deciding who advances. The system is sold to EU employers.
Use the deterministic Article 6 / Annex III decision tree. Do not assume missing facts.
Generate
- the likely classification;
- the facts and legal rules supporting it;
- unresolved legal questions;
- the high-risk obligations readiness checklist;
- Annex IV documentation gaps; and
- an Annex VIII registration-data draft.
Clearly state the legal baseline date and do not describe the result as legal certification.
- Read your context and instructions
- Compiled the eu ai act
EU AI Act pre-assessment
Likely classification: HIGH_RISK_ANNEX_III
Annex III area: Employment, worker management and access to self-employment.
Confidence: High on the Annex III use-case match; legal review remains recommended for final product-specific conclusions.
Why
The documented intended purpose is to rank and recommend job applicants before a recruiter decides who advances.
That use case matches the Annex III employment category covering AI intended for recruitment or selection, including application analysis, filtering, and candidate evaluation.
The documentation also indicates that the AI score is presented before the recruiter's advancement decision, so the system appears to materially influence a decision affecting a natural person.
No Article 6(3) exception has been applied because the available documentation does not establish that the tool is merely procedural or that it does not materially influence the decision outcome.
Current timeline
The legal baseline for this assessment is the consolidated EU AI Act following the July 2026 amendments.
Annex III high-risk rules are scheduled to apply from 2 December 2027.
This assessment is therefore a readiness review for those requirements, not a statement that every high-risk obligation is already enforceable today.
Readiness snapshot
| Area | Status |
|---|---|
| Risk-management system | PARTIAL |
| Data governance | PARTIAL |
| Technical documentation | PARTIAL |
| Automatic logging | PRESENT |
| Human oversight | PARTIAL |
| Accuracy / robustness | PARTIAL |
| Cybersecurity | MISSING |
| Quality-management evidence | MISSING |
| Conformity-assessment route | NOT_ASSESSED |
| EU database registration data | PARTIAL |
Key gaps
Human oversight
The supplied documentation states that recruiters can override recommendations, but does not explain how reviewers are trained to identify automation bias or when escalation is required.
Data governance
The model manifest identifies training sources but does not document representativeness testing for the EU employment context.
Cybersecurity
No system-specific robustness or adversarial testing evidence was found.
Annex VIII preparation
A draft registration-data file was generated from known facts.
It is explicitly marked as a
preparation artifact only and is not represented as an official EU database JSON submission schema.
Legal review questions
- Does the candidate recommendation constitute profiling of natural persons?
- Which entity is the AI Act provider where customers configure ranking criteria?
- Could customer-specific modifications constitute a substantial modification?
- What conformity-assessment route applies to the final product?
- Do particular deployers trigger a fundamental-rights impact assessment under Article 27?
Files generated
- classification.json
- classification_report.md
- obligations_checklist.md
- annex_iv_gap_analysis.md
- annex_viii_registration_draft.json
- decision_log.json
Important: This is a structured legal-operations pre-assessment, not legal advice or certification.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
EU AI Act classification is not a simple keyword exercise. Teams must determine the system's intended purpose, test Annex I and Annex III routes, assess the Article 6(3) exception, identify profiling, and map the result to technical-documentation and governance requirements.
A wrong classification can create serious compliance exposure—or cause teams to spend months implementing controls that do not apply.
What it does
EU AI Act High-Risk Assessor converts product, model, and architecture documentation into a structured compliance pre-assessment.
- Applies a deterministic Article 6 decision tree
- Checks all eight Annex III high-risk areas
- Evaluates the Article 6(3) exception separately
- Applies the profiling override for Annex III systems
- Flags Annex I product-law questions for specialist review
- Produces an explainable classification memo
- Maps high-risk requirements and governance obligations
- Creates an Annex IV technical-documentation gap analysis
- Prepares an Annex VIII registration-data draft
- Records unresolved assumptions for legal review
Why this beats prompting it yourself
A generic legal prompt can drift between rules, omit exceptions, or confidently fill factual gaps.
This skill forces every assessment through a versioned legal decision tree and only permits defined classification outcomes. Missing facts remain missing, while ambiguous product-law questions are escalated instead of guessed.
Built for the current EU AI Act timeline
The package reflects the 2026 amended implementation timeline: Annex III high-risk requirements apply from 2 December 2027, while regulated-product high-risk rules apply from 2 August 2028.
Designed for legal review
This is a compliance-readiness tool, not legal certification.
It does not replace qualified EU counsel or a conformity assessment body, and it does not submit information directly to the EU database.
The generated Annex VIII JSON is a preparation artifact—not an official EU submission schema.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Legal & Compliance.
Requires technical documentation (PRDs, model cards, or architecture logs) to extract facts. Assessment reflects the July 2026 legal baseline. Does not submit data to the EU database.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 months ago
- Software
Python