Deep Eye — ИИ-пентест веб-приложений

★ 2.2K

zakirkun/deep-eye is an open-source AI-driven penetration testing tool for web applications, written in Python 3.8+ and run via CLI. Reach for it when you need to audit your own app before shipping: sweep an API across 50+ vulnerability classes (SQLi, XSS, SSRF, JWT, IDOR, GraphQL, CORS/CSP, supply-chain JS, and more), generate a compliance-mapped report for PCI-DSS v4, SOC2, or ISO 27001, or supercharge manual pentesting with context-aware AI-generated payloads. Under the hood it orchestrates 13+ AI providers — OpenAI, Claude, Grok, Gemini, Ollama, Mistral, Groq, LM Studio, LiteLLM, and others — with automatic failover, a RAG-indexed CVE database from NVD/MITRE/Exploit-DB, Playwright browser automation, mitmproxy integration, OpenAPI/Swagger spec ingestion, finding deduplication, and export to HTML, PDF, JSON, SARIF, JUnit, CSV, and XLSX. Running local models via Ollama keeps all data on-premises. This is a dual-use tool — use it only against systems you own or have explicit written authorisation to test; scanning third-party targets without consent is illegal. Compared to OWASP ZAP, payloads are generated dynamically against the detected stack and CVE context rather than pulled f