PURE CORE
Deterministic VBA
Explicit inputs and immutable configuration. Planned core procedures depend on their arguments, without reading worksheets, files or mutable host state.
Read the VBA contracts ↗Liquidity risk / Reference and independent validation
A clearer view
of liquidity.
Built for scrutiny.
An Excel/VBA foundation for a traceable Liquidity Coverage Ratio reference.
Separate planned EU and Basel profiles. Explicit inputs, reviewable rules and independent evidence, from the source decision to the reported result.
Distinct EU and Basel profilesVBA source contractsSynthetic data onlyMPL-2.0 licensed
01 / Scope
One project. Two explicit profiles.
K-LCR is being prepared for people who need to inspect and independently validate a Liquidity Coverage Ratio calculation. The foundation makes the intended method, its inputs and the evidence behind each result visible.
- 01 / DEFINE
Identify the applicable profile
EU and Basel are distinct planned profiles. Source editions, reporting dates and parameter versions must be explicit before a calculation can be implemented.
- 02 / TRACE
Follow inputs through the rules
The planned scope covers liquidity-buffer classification, inflows, outflows, intermediate reconciliations and ratio outputs. Exact eligibility and treatment rules remain to be specified.
- 03 / VERIFY
Build independent evidence
Use synthetic cases, independently derived expected values and a real Excel host run. Each claim must point to its source, tested revision and recorded result.
Production calculations, regulatory reporting submission, real portfolio data and numerical certification are outside the repository-setup milestone.
02 / Architecture
Keep the method visible.
Keep the layers deliberate.
The selected stack is Excel/VBA. The architecture separates deterministic calculation from request validation and Excel interaction. Python and Node support repository checks; they are not an alternative production calculation engine.
PUBLIC API
Validate the whole request
Deliberate entry points, named errors and documented types. Keep profile and version selection explicit, and preserve caller-owned inputs.
Explore the boundaries ↗EXCEL HOST
Controlled input and output
The planned adapter owns workbook interaction and state restoration. A macro-free template, source synchronization and host compatibility still require implementation and evidence.
Read the host target ↗03 / Get started
Begin with the source and its contracts.
Inspect the active development branch
Clone the public source and read the prerequisites before running the portable checks.
git clone --branch release/0.1.0 https://github.com/danielep71/K-LCR.git
cd K-LCR
python tools/check.py --ciThe command records the candidate and check results. It requires the tooling described in the installation guide. There is currently no workbook to download or install.
Open developer setup ↗04 / Evidence and boundaries
Make every claim inspectable.
AVAILABLE NOW
Portable repository checks
Repository integrity checks tooling, configuration, tracked text, Python quality and workflow syntax. A passing run applies to its candidate and scope.
Inspect the workflow runs ↗PLANNED HOST EVIDENCE
Import. Compile. Run.
No Excel host run has been recorded. A genuine smoke test must identify the candidate, environment, cases and cleanup result before a host claim can be made.
Follow the host-evidence work ↗PLANNED NUMERICAL EVIDENCE
Independent expectations
Regulatory sources, field-level contracts, independent fixtures and tolerances remain setup work. Static checks alone cannot establish numerical correctness.
Follow source registration ↗Follow validation contracts ↗Reference / Research / Independent validation