What problem does it solve?
It helps detect inconsistencies and incorrect automatic classifications in the CABYS–CIIU affinity logic and the monthly cut workflow, so a firm can trust which items are marked for purchases versus expenses and when manual decisions are required.
Core Features & Use Cases
- CABYS–CIIU affinity consistency auditing: Checks how the system decides COMPRAS vs GASTOS using CABYS “bien_es_compra|is_purchase”, forced purchase chapters (“compras_capitulos|purchase_chapters”), and CIIU sector rules (“ciiu_affinity|affinity”).
- Ambiguity escalations review: Verifies the actual criteria for items flagged as ambiguous and routed to the manual decision modal (“gestor_contable/gui/corte_ambiguo_modal.py”).
- Vendor catalog contamination and decision persistence checks: Audits whether vendor catalog inputs and saved manual decisions (“vendor_catalog|VendorCatalog”, “save_decision|persist|guardar”) can bias outcomes across customers/periods.
- Monthly cut execution sanity: Reviews the monthly cut flow for pending/undecided items, batch processing, and whether rollback/reversibility is implemented as intended.
Quick Start
Ask the agent to audit the current CABYS–CIIU and monthly cut code paths and produce an evidence-based report including files reviewed, pending-decision handling, and any detected problems with quoted code excerpts.