ghost-scan-code

ghost-scan-code is a static application security testing (SAST) skill designed to identify security vulnerabilities in source code repositories. It helps developers and security teams audit applications and libraries by planning and executing targeted scans against common vulnerability categories such as SQL injection, cross-site scripting (XSS), broken object level authorization (BOLA), broken function level authorization (BFLA), server-side request forgery (SSRF), unsafe deserialization, prototype pollution, path traversal, ReDoS, and zip slip issues. The skill supports multiple project types including backend, frontend, mobile applications, and reusable libraries, making it suitable for broad code security assessments across modern software stacks.

The skill includes a structured workflow that prepares repository-specific scan directories, loads scan criteria mappings, generates project scan plans, and nominates candidate files for analysis. It supports configurable scan depth modes including quick, balanced, and full scans, allowing users to trade off between speed and coverage. The workflow also incorporates resumable execution behavior for long-running scans and validates generated planning artifacts before continuing. By leveraging scripts, criteria definitions, and loop-based execution, the skill organizes vulnerability scanning tasks into repeatable and auditable stages.

This skill is intended for developers, application security engineers, DevSecOps teams, and auditors who need to perform source code security reviews or vulnerability discovery in repositories. It is useful during secure development reviews, pre-release security validation, CI/CD security workflows, dependency and library audits, and exploratory vulnerability assessments of unfamiliar codebases. Users can also customize scan behavior through arguments that focus on specific vectors, candidate files, or project areas.