dkl101001/bloomcore-basics

BLOOMCORE Basics is the public-facing piece of a larger project called BLOOMCORE that its author has not released in full. The README is careful to say the same thing several times: this repository is not the complete system. The private parts, which the author refers to as the synthesis logic, orchestration, scoring, contradiction handling, and full assembly, stay separate. What is shared here is the public surface, meant to be inspectable and reusable on its own. The central idea is something the README calls receipt-native, audit-aware system design. The pattern works like this: when an event happens inside a system, a structured record called a receipt is produced. The receipt describes what happened, what state was involved, what hash anchors it, and what previous event it links to. Validators check the receipt against a schema, hashes are confirmed, and a readout is generated, leaving an inspectable audit trail. The point is that the engine running the system can be non-deterministic, with mutation and exploration and stochastic choices, while the record of what happened stays deterministic and replayable. The public layer in this repo contains JSON schemas for receipts, public node readouts, and validation results, plus example files for each. It also sketches a small Python package called bloomcore_basics with utilities for canonical JSON serialization, hash helpers, and receipt and readout handling. The README shows a suggested directory layout under schemas/, examples/, and src/bloomcore_basics/, and points to a minimal_run/ folder for a local demonstration. The license is listed as Apache 2.0 or MPL 2.0. The repository has 0 stars and the language tag is unset, which is consistent with a freshly seeded public scaffold rather than an active codebase. The README itself reads more like a position statement and a vocabulary document than a guide to running working code.