Security Research · Detection Engineering · Validation Tooling
I investigate application, cloud, CI/CD, and smart-contract security boundaries. My work centers on deterministic reproduction, attacker-capability analysis, evidence preservation, and responsible disclosure.
I build tooling that helps researchers move from an interesting signal to a defensible conclusion—without confusing scanner output, anomalous behavior, or theoretical risk with demonstrated security impact.
- Casper — deterministic validation and evidence runtime for security research. Casper evaluates candidate findings, evidence sufficiency, reproducibility, and advancement decisions; it is intentionally not a vulnerability scanner.
- SQUIRM — scope-aware endpoint collection and entropy-assisted triage for authorized security assessment.
- shhh-ai — AI-assisted secret-candidate triage with redaction-by-default output and conservative LLM review.
A finding should state:
- the security boundary being crossed;
- the minimum attacker capability;
- a deterministic reproduction path;
- observable impact supported by preserved evidence; and
- the authorization and scope under which validation occurred.
Uncertainty stays visible. AI review is advisory. Impact is never inferred from a tool label alone.
Detection engineering · application security · cloud and CI/CD security · smart-contract review · security automation · evidence-driven vulnerability validation
All security research represented here is performed on systems I own or under explicit authorization and applicable program rules.