Rust systems · Agent infrastructure · Applied LLM research
Reproducible experiments, explicit boundaries, measured behavior.
Durable, bounded execution for coding-agent experiments with explicit verification and constrained tool authority.
README / Quick Start · Architecture · Evaluation Evidence
Evidence-first repository retrieval in Rust with lexical, neural and hybrid search plus extractive source selection.
Architecture · Results · Reproducibility
Controlled infrastructure for iterative model-guided candidate generation, validation, evaluation and selection. No general RSI or intelligence-improvement claim is made.
Looped-transformer experiments with bounded RSI-style iterative adaptation and post-selection held-out evaluation.
Reproducible model-adaptation experiments on Rust compiler-error explanations, including preserved negative results.
- grainx — terminal system monitor with an optional loopback HTTP metrics service and JSON/CSV export.
- deltasafe — authenticated file transfer for a trusted LAN, with bounded frames and verified file publication.
- RustHound — streaming log-analysis CLI with configurable pattern, frequency, and correlation rules.
Experiments are documented with committed artifacts and reproducible commands, and system boundaries are stated explicitly rather than left implied. Negative and inconclusive results stay in the record when they are part of the evidence.