What problem does it solve? Manually auditing a Python codebase for vulnerabilities is slow and error-prone, and raw SAST scanner output is noisy with false positives. This Skill automates the entire audit: it installs and verifies PySpector (including its required Rust toolchain), selects the right scan flags based on what the codebase actually contains, runs the scan, and statically re-verifies each finding against the real source before reporting only confirmed issues. ## Core Features & Use Cases - End-to-end audit workflow: Handles installation, Rust toolchain setup, CLI flag discovery via live --help output, codebase recon, scanning, verification, and reporting. - Context-aware flag selection: Detects LLM/AI library usage (openai, langchain, etc.) to enable AI-specific rules, and assesses dependency pinning to decide on supply-chain CVE checks against OSV.dev. - False-positive filtering: Re-reads flagged files and traces taint paths from source to sink, classifying findings as confirmed true positives, false positives, or needing human judgement. - Use Case: Point the agent at a FastAPI project and ask for a security audit; it installs PySpector, scans with the appropriate rulesets, verifies each finding against the code, and delivers an HTML report plus a summary of confirmed vulnerabilities with file, line, and exploitability reasoning. ## Quick Start Ask the agent to run a security audit of your Python project with PySpector and report only verified vulnerabilities.
