Broken Numbering Detection

Capabilities { .capabilities }

Design a solution that detects all signs of broken automatic list numbering in a legal contract. The solution must identify at least four distinct patterns of numbering errors: sub-clause numbering that resets mid-document, duplicate clause numbers, skipped clause numbers, and empty list items creating phantom clauses.

• Given a document where sub-clauses reset to "1.1" inside a higher-numbered section (e.g., "1.1" appearing within clause 10), the solution flags this as a numbering reset error @test
• Given a document with two clauses both numbered "26.3", the solution flags the duplicate clause number @test
• Given a document that jumps from "26.2" to "26.4" with no "26.3", the solution flags the skipped number @test
• Given a document with an empty list item that creates a phantom clause number, the solution flags it for removal @test

Implementation { .implementation }

The solution should operate on a clause map and the raw document content. It should perform four distinct checks:

1. Reset detection: scan for sub-clause numbers that revert to a low value (e.g., 1.1) in a position where the parent clause number is much higher, indicating an automatic numbering reset.
2. Duplicate detection: compare all clause numbers in the map and flag any number that appears more than once.
3. Skip detection: for each parent clause, verify that its sub-clauses form a contiguous sequence with no gaps; flag any missing number.
4. Phantom clause detection: identify empty or blank list items in the document that have been assigned a clause number by the automatic numbering system, and flag them for removal.

Each detected error should include the error type and its location within the document.

API { .api }

Accepts a clause map and raw document content. Returns a list of detected numbering errors, where each error contains:

• type: one of reset, duplicate, skip, or phantom
• location: the clause number or position in the document where the error was found

Dependencies { .dependencies }

contract-numbering-fixer 0.1.0 { .dependency }

AI skill that reviews and fixes clause numbering issues and stale cross-references in legal contracts.