Applies to: Machinery, partly completed machinery, related products placed on the EU market
Regulation (EU) 2023/1230 (the "Machinery Regulation") replaces the 2006 Machinery Directive on 20 January 2027. It explicitly adds cybersecurity essential health and safety requirements (EHSRs), brings AI-system safety and self-evolving machinery into scope, and modernises conformity assessment for software-updated and connected products. For robotics teams shipping into the EU, this is the proximate compliance forcing function.
Audit-readiness, not certified. Roboticks is not itself a certified verification tool. Conformity assessment is the manufacturer's responsibility (with notified-body involvement where required). We assemble the verification evidence your assessor reads.
Annex III adds new EHSRs covering: (1) protection against corruption of safety functions (cybersecurity); (2) safety of machinery embedding AI systems; (3) information for safe use including digital instructions; (4) traceability and identifiability for software components. Conformity assessment under Annex IV explicitly covers connected and learning machinery. Substantial modifications trigger re-conformity.
Manufacturers, importers, distributors, and authorised representatives placing machinery on the EU market. Industrial robots and cobots (mostly), AMRs and AGVs (mostly via being machinery), autonomous mobile platforms (yes when not type-approved as vehicles), surgical robots (separate medical device regs apply primarily but machinery aspects in scope).
Cybersecurity EHSRs become explicit. AI systems in safety functions get a dedicated provision. Digital instructions for use are explicitly allowed. Substantial modifications (including via software update) require re-assessment. Annex I high-risk list now includes machinery with self-evolving behavior and machinery embedding AI safety functions.
Ingest your requirements (often derived from EN 60204-1, ISO 12100, ISO 10218, ISO 13849, and the regulation's Annex III). Link tests via @confirms; the matrix shows live coverage. Evidence pack at release time bundles JUnit + SARIF + ReqIF + MCAP into a hash-chained ZIP that the conformity assessor can verify. We are not a certified tool; we assemble the evidence the assessor reads.
These are example requirements you could derive from EU MR 2023/1230 and link to verifying tests via the @confirms decorator. Not exhaustive and not a substitute for a domain-expert risk assessment.
EU-MR-1.2.1Control systems must be designed so that fault conditions do not lead to dangerous situations. Verifiable via fault-injection tests on safety functions (e.g. emergency stop reachability under simulated bus disconnection).
EU-MR-1.1.9Cybersecurity essential health and safety requirement. Verifiable via SARIF static analysis, threat modeling artifacts, SBOM, and tests covering message authentication on safety-critical topics.
EU-MR-1.1.6Machinery embedding AI systems intended to safeguard safety functions must be verifiable, with documentation of the AI's behavior. Typically verifiable via sim-in-the-loop scenario coverage plus inference-output regression tests.
Ingest EU MR 2023/1230 as a ReqIF export from your requirements tool, or as a PDF for the LLM extractor. Link tests via @confirms. Watch the matrix turn green as PRs land.