GAIA is an on-premise static analysis tool that detects exposed secrets and insecure patterns in source code. It is designed for environments where sending source code to a third-party cloud service is unacceptable: regulated FinTech, healthcare, defense, and any air-gapped pipeline.
The tool is a single native binary (Linux x86-64). It has no runtime dependencies, no installer, and no configuration server.
GAIA performs no network activity at any point — not during scanning, not during license verification. This is not a policy setting; it is a property of the build.
Two languages, each doing what it is best at:
| Layer | Language | Responsibility |
|---|---|---|
| Parsing & I/O | Rust (safe) | Directory traversal, pattern extraction, report generation. Memory-safe, no manual allocation. |
| Rule & risk core | Ada / SPARK | Severity classification and financial-risk scoring, compiled statically into the binary (FFI). Formally verified with SPARK (gnatprove): proven free of runtime errors — no overflow, no range violation, guaranteed termination. The scoring IP is not shipped as readable source. |
Detection is fully deterministic — the same input always yields the same output. There is no AI, no model, no probabilistic component. Results are reproducible and auditable.
The Ada rule core carries SPARK contracts (preconditions, postconditions) and passes gnatprove with all checks proved — a machine-checked mathematical guarantee that the risk-scoring logic cannot crash or overflow on any input.
Licenses are Ed25519-signed tokens. Each key carries a product, tier and identifier, signed with a private key held only by the vendor. The binary embeds the corresponding public key and verifies signatures locally.
Before release, GAIA is validated beyond unit tests:
Memory safety is provided by the Rust layer; the Ada rule core is pure, allocation-free arithmetic, which is structurally leak-free.
Because GAIA never transmits data off the host, it introduces no new data-processing surface. It supports data-residency and sovereignty requirements under GDPR and KVKK by keeping all source, findings and telemetry local. GAIA is a tool that aids compliance; it is not itself a certification.
Drop the binary into a CI/CD stage or run it on a developer workstation:
For unlimited-developer, air-gapped, or procurement-based deployments, see Team / Site licensing or contact webacan@gmail.com.