Adding GGUF support without mislabeling SLM1

A concrete parser, tokenizer, tensor, quantization, test, and compatibility plan for real GGUF support alongside the existing custom container.

The inspected runtime implements custom SLM1. True GGUF compatibility requires a separate metadata and tensor parser, supported-architecture adapters, tokenizer import, named quantization kernels or a controlled conversion step, malformed-file tests, and differential logits.

Format honesty

SLM1 remains an explicit project-owned format. A project or domain name cannot substitute for a loader contract.

Compatibility evidence

Publish exact artifact hashes, GGUF version and architecture, tokenizer, quantization type, runtime/WASM hash, reference outputs, browser/device matrix, and limitations.