Technical file skeleton

The problem this solves

The technical file is usually assembled once, from whatever the folders hold the week before CE marking, and its order and completeness are discovered at the same moment an assessor discovers them. Compiled from the model, the file's structure exists from day one: every section is either generated, stubbed with the exact external artifact that belongs there, or marked authored, and the completeness table keeps the score in the open.

How it works

The binder in Annex IV order: model-derived sections filled, the rest stubbed, a completeness table on page one. Runs the risk assessment documentation, the machine description and the evidence index as sections of one document, inserts reference stubs for the sections that live outside the model, and shows an assessor, and the team, what is done and what is not.

Run the audits first

Regulation (EU) 2023/1230 applies from 20 January 2027, and the technical file is kept for ten years after the machine is placed on the market. This workflow compiles a document from the model; it is not a conformity assessment, and it assumes all five audits have been run and their findings acted on.

1. 01Run the Annex III requirement gap audit first and close the gaps it finds, so the file is compiled from a model that carries the requirements the law asks for.
2. 02Run the Annex IV readiness audit and have the team confirm the Annex III clause list: the confirmed list is this workflow's input.
3. 03Run the hazard register integrity audit, so the item (b) section is built on a register that holds together.
4. 04Run the verification coverage audit, so the test reports section indexes evidence rather than gaps.
5. 05Run the architecture and interface completeness audit, so the description and operating sections have no holes to inherit.

Compile the technical file skeleton under Regulation (EU) 2023/1230, Annex IV Part A, from the connected Dalus model. Do not modify the model.

Input: the confirmed list of applicable Annex III clauses, pasted below the prompt. If absent, propose and mark unconfirmed.

Step 1. Read the entire model.

Step 2. Completeness table on page one: one row per Annex IV Part A item in the Regulation's order and letters. Columns: item, requirement in one line, source (model, external system, authored), status (complete, partial, authored, external), and the section where it appears.

Step 3. Sections, in the Regulation's order:
(a) and (d): generate exactly as the "Machine description and operating description" workflow.
(b): generate exactly as the "Risk assessment documentation" workflow.
(c): a reference section listing every part with an Onshape link or linked document, and stubs for mechanical drawings, electrical schematics and circuit diagrams held elsewhere. Do not describe drawings the model does not reference.
(e): a reference section listing standards cited by requirements and test cases, by number only, with the note that the team must confirm editions against the Official Journal. No currency judgement.
Test reports: generate exactly as the "Verification evidence index" workflow.
Instructions for use: a stub, plus the residual risk summary from (b) as the input for the residual risks section of the instructions.
Declaration of conformity and series-production measures: stubs.

Step 4. Combined gap annex: merge the gap annexes of (a)/(d), (b) and the evidence index, deduplicate, rank safety gaps first.

Step 5. Combined provenance annex.

Output rules:
- One Word document titled "Technical Documentation Skeleton, Regulation (EU) 2023/1230 Annex IV Part A", title block with model name, date and revision placeholder, completeness table on page one.
- Regulation's item letters as section headings.
- Stubs state exactly which external artifact belongs there. No invented content anywhere.

What you get

Common questions