euaiact_annex_iv_checklist
ClaudeThe nine titles and descriptions summarise Annex IV. sub_items are non-binding implementation prompts, are not verbatim Annex IV text, and do not create additional legal requirements. Under Article 11(1), SMEs, including start-ups, and SMCs may provide the Annex IV elements in a simplified manner only by using the Commission form referred to in that paragraph.
euaiact_answer_question
ClaudeSearch frequently asked questions about the EU AI Act and get best-match answers with article references. Covers classification, deadlines, roles, governance, documentation, risk assessment, penalties, GPAI systemic risk, FRIA, transparency, and sector-specific guidance.
euaiact_assess_art6_3_exception
ClaudeWalk through the Art. 6(3) exception for Annex III high-risk AI systems. An Annex III system is NOT high-risk only if the provider affirmatively assesses no significant risk to health, safety, or fundamental rights, AND the system falls under one of the four conditions: (a) narrow procedural task, (b) improves prior human activity, (c) detects patterns without replacing human assessment, (d) preparatory task. Set no_significant_risk_to_health_safety_fundamental_rights=true only when that threshold has been assessed. CRITICAL: The exception does NOT apply if the system performs profiling of natural persons (Art. 6(3), third subparagraph). Providers invoking the exception must document the assessment (Art. 6(4)) and still register in the EU database (Art. 49(2)).
euaiact_assess_system
ClaudeAssess one normalized EU AI Act system profile against the pinned sealed corpus. Returns separate legal-classification, impact, and implementation-readiness blocks with field-level facts, decisive missing facts, complete finding provenance, warnings, and recommended atomic follow-up calls. Sparse inputs fail closed. Impact never changes legal classification, and classification never implies implementation readiness. HOW TO BUILD THE PROFILE: every fact is an object {fact_id, value, origin, verification, evidence_reference_ids}. For a fact the user told you, use origin "explicit_structured_input", verification "caller_asserted", evidence_reference_ids [] and any unique lowercase fact_id such as "identity.system_name". A useful minimum is identity.system_name, intended_use.intended_purpose (intended_use also needs reasonably_foreseeable_uses, which may be []), role_facts.roles, geography with jurisdictions, affected_person_groups and one of used_in_eu, placed_on_eu_market or output_used_in_eu, and decision_context.decision_consequence. Set annex_iii.domain and annex_iii.annex_iii_point only when the system itself performs a use Annex III lists for that area, never because it operates in that sector. Supply only facts you were given: a missing decisive fact comes back in missing_facts with the question to ask.
euaiact_calculate_penalty
ClaudeCalculates the maximum possible fine for an EU AI Act violation based on violation type, global annual turnover, SME status and SMC (small mid-cap) status. Implements the Art. 99 penalty framework including the SME/startup lower-of rule (Art. 99(6), tiers 99(3)-(5)) and the narrower SMC rule (Art. 99(6a), tiers 99(4)-(5) only; no SMC cap on Art. 5 fines and none under Art. 101). Returns a comparative block so the agent can show the SME reduction to the user.
euaiact_check_deadlines
ClaudeReturns key implementation milestones and deadlines for the EU AI Act with days remaining, a next_milestone shortcut, and a summary of the Digital Omnibus. The milestone timeline always reflects the operative law. Set include_pending_omnibus: true to also receive the structured Digital Omnibus pack (each item source-status labelled; its status and enacted fields carry the current legislative state). Use only_upcoming: true to drop past milestones.
euaiact_check_gpai_systemic_risk
ClaudeDetermine whether a general-purpose AI model qualifies as a GPAI model with systemic risk under Art. 51. This tool uses an adjudicated conservative boundary at or above 10^25 cumulative training FLOPs for the Art. 51(2) presumption. The Commission may also designate models with equivalent capabilities or impact under Art. 51(1)(b). Returns baseline GPAI obligations under Art. 53 plus systemic-risk-only obligations under Art. 55, and the Art. 52 notification duty.
euaiact_classify_system
ClaudeClassify an AI system's risk level under the EU AI Act (Regulation 2024/1689). HOW TO CALL: derive the structured signals from the user's description yourself and pass them together with description. The result is decided deterministically on the signals; free text alone is only matched against keywords and often returns insufficient_information. Set a signal only when the description supports it and leave it out otherwise; never guess, and never infer signals.domain from the sector a system operates in (read that field's description). missing_signals lists unprovided fields, including optional context; it is not a checklist of facts required for the reported result. Relay next_questions where present, then call again with the answers. A matched route does not establish that every other route has been ruled out. Returns risk classification, applicable Annex III category, relevant articles, provider/deployer determination, matched signals, and follow-up questions the agent should relay. Note: Art. 6(3) exceptions require documented justification and cannot be auto-applied; use euaiact_assess_art6_3_exception.
euaiact_get_article
ClaudeRetrieve an operational summary of a specific article of the EU AI Act (Regulation 2024/1689), plus a stable EUR-Lex URL to the canonical text. Supports a subset of the most-cited articles (Art. 3, 4, 5, 6, 9-17, 26, 27, 43, 47, 49, 50, 51, 53, 55, 72, 73, 99, 100, 113). For articles outside this subset the tool returns the EUR-Lex base URL. The summary is not statutory text: follow the URL and verify the official provision before quoting it.
euaiact_get_obligations
ClaudeReturns specific compliance obligations for providers or deployers based on AI system risk level.
