With no source code and no API documentation, the analyst is limited to black-box techniques, and fuzzing is the canonical way to find input-validation flaws by bombarding the application's inputs with malformed, unexpected, or boundary data and watching for failures. SAST is ruled out because it requires source code, and agentless scanning is a configuration/vulnerability-scan technique rather than targeted input testing. Reverse engineering is far more effort and is aimed at understanding internals, not systematically exercising input handling, so fuzzing is the best fit.
Community Discussion