#1LLM security, prompt injection, model extraction, adversarial AI, and AI red teaming tools.
Kitploit recommended

CyberRAG — RAG 100% local (Ollama) sobre apuntes y libros de ciberseguridad: recuperación híbrida, respuestas con citas, asistente de pentest por…
A Proof-of-concept repository showing how an untrusted MCP server can steal literally everything...

Exploits for CVE-2024-14027 creating using AI/Claude as a test.

Research pipeline for detecting latent indirect prompt-injection exposure signals in agentic LLMs via hidden-state probing, including trace…

Research code reproducing multi-turn LLM jailbreak experiments (FITD, MRCJ, ActorAttack, X-Teaming) from the SoK intent-oriented systematization…

Static analysis CLI that scans codebases for LLM prompt-injection, data-exfiltration, jailbreak, and unsafe agent/tool vulnerabilities. Runs fully…

Expose and detail an unauthenticated stored XSS vulnerability in the Google Cloud Vertex AI Python SDK affecting versions 1.98.0 to 1.130.9.

Detects LLM context-leakage attacks by training lightweight behavior probes on log-probabilities, with vLLM offline/server detection pipelines.

OWASP project for AI-assisted threat modeling, helping teams identify, analyze, and document security threats and mitigations in system designs.

Pre-execution action-auditing defense that detects and masks indirect prompt injection in tool-using LLM agents using embedding retrieval and…

Browser-local security monorepo with six modules for mobile APK/IPA triage, client-side DAST fuzzing, OSINT directories, offline AI threat scoring,…

Jailbreak vulnerability across early-exit depths LLMs. Published at NeurIPS 2026 in the FLMSec workshop

Zero-trust sandbox for AI agents with kernel-level filesystem jail, transparent network proxy, and YAML-based policy engine to intercept and control…

Agentic harness that automates vulnerability discovery and validation, orchestrating AI-driven analysis workflows to confirm exploitable findings.

Deterministic benchmark and evaluation harness for persistent-memory poisoning in LLM agents, with paired clean/poisoned cases, LangChain/LlamaIndex…

Local test range for agent memory poisoning - watch a planted memory steer an agent, then prove the repair. Synthetic targets, scripted deterministic…

Open reference implementation of the VEP protocol (RFC-010): a deterministic sandbox and benchmark suite evaluating AI agent runtime containment…

LLM-driven agentic group shilling attack framework that manipulates black-box collaborative-filtering recommender rankings using adaptive multi-role…