code-understanding

To trace data flows for vulnerability hunting, you follow untrusted inputs from entry points through the full call chain to dangerous sinks. This builds ground-truth understanding and generates structured flow traces with evidence for validation.

Mapping an attack surface involves building a ground-truth context model of entry points, trust boundaries, and dangerous sinks from an attacker's perspective. This provides structured context maps before or alongside static analysis.

Yes, you can apply code comprehension alongside static analysis to build ground-truth understanding. It maps architecture, traces data paths, and hunts vulnerability variants to validate and contextualize static analysis findings.

To find vulnerability variants, you perform a systematic variant hunt across the codebase and group results by root cause. This generates structured variant reports that scope fixes for the identified vulnerability patterns.

Analyzing unfamiliar code for secure analysis requires multiple modes: map for architecture, trace for data flows, hunt for vulnerability variants, and teach for in-depth pattern explanation. These output structured formats with clear validation hooks.

You should use code comprehension before or alongside static analysis rather than replacing it, as it requires structured outputs and validation hooks to build ground-truth understanding of entry points, data flows, and vulnerability patterns.