Before you start
Named tender documents, revisions and dates, trade codes and available measurement evidence. The session must support reading your files and creating the requested format. Pack deliverables are in French.
DCE document consistency check with ChatGPT: method →
1. Download the skill
Download the ZIP pack below and keep it without extracting it. The standalone SKILL.md file is also available under details, with the exact filename required by the importer.
Download the ZIP pack ↓2. Import the skill into ChatGPT
On the skill import screen, select the downloaded ZIP and complete the import. Then open a conversation and ask ChatGPT to use the “controle-coherence-dce” skill. If your interface does not offer skill imports, paste the contents of SKILL.md into the conversation as instructions.
3. Add project documents
Named tender documents, revisions and dates, trade codes and available measurement evidence.
4. Generate and review the draft
Inventory revisions first. Map trade, element, CCTP article, DPGF item and drawing reference. For each discrepancy request two precise references, a possible consequence and a proposed correction. Separate confirmed anomalies from unverifiable points.
A request to try
Use my installed skill « controle-coherence-dce ». Fictional example: drawing P02 revision B shows a 90 cm door while CCTP article 4.2 specifies 83 cm. Record a referenced discrepancy; do not pick the correct dimension or change source files. Create a downloadable French draft in XLSX. If a capability is unavailable, say so without claiming to have created a file.
5. Check the result
- Review both sources for each discrepancy and actual review coverage. Finding no discrepancies does not show that all documents were checked.
- Open the file and check readability, incomplete fields and source references.