replica-rmc

Automating an RMC replica involves orchestrating PDF fatiation, deterministic facts extraction, and DOCX drafting to transform a process folder into a ready-to-review document. The skill handles end-to-end generation, applying rule-conformant formatting like Cambria font.

Deterministic facts extraction creates a _facts.json source of truth from process PDFs, which is then used for evidence validation. Validate_against_facts verifies the generated replica text against these extracted facts and original PDF text to identify critical points.

Yes, this legal drafting workflow supports process PDFs or folders specifically for AM, AL, BA, and MG states. It applies tailored editorial rules and bank scenario checks to generate the RMC/RCC replica for these jurisdictions.

To format a legal replica in DOCX, the skill applies rule-conformant formatting constraints such as mandatory Cambria font, specific structure requirements, and list checks. It outputs a fully formatted document ready for review.

Yes, the workflow utilizes pymupdf and fitz for PDF processing and fatiation, alongside python-docx for drafting the final replica. These dependencies enable seamless extraction and document generation from consolidated process folders.

Automated extraction requires a mandatory SKILL entrypoint with YAML frontmatter and relies on deterministic extract_facts output as the source of truth. Limitations include dependency on accurate PDF text and the need for manual review of critical points identified during validation.