cargo-fuzz

cargo fuzz Acargo subcommand for fuzzing with libFuzzer! Easy to use! Installation $ cargo install cargo-fuzz Note: libFuzzer needs LLVM sanitizer support, so this only works on x86-64 and Aarch64, and only on Unix-like operating systems not Windows. This also needs a nightly compiler since it us...

CVE-2026-84286

The exllamav3_ext CUDA extension is vulnerable to an out-of-bounds (OOB) memory access within the kernel-table dispatch process. The root cause is a missing bounds check where a model-supplied index—specifically the checkpoint-derived block index (cbi ) and kernel parameter K —is used to access a...

atomic-red-team

Atomic Red Team Atomic Red Team™ es una biblioteca de pruebas asociadas al framework MITRE ATT&CK®. Los equipos de seguridad pueden usar Atomic Red Team para probar sus entornos de manera rápida, portable y reproducible. Comenzar Puede ejecutar pruebas atómicas directamente desde la línea de...

security-operations-labs

Security Operations Labs Overview This repository consolidates defensive security work from a university context in a form reworked for public traceability. The focus is on methodology, reproducible labs, validation, and the limitations of the measures developed—not on publishing course material...

turnstile

Archivo de investigación de Turnstile Código de investigación histórico y notas técnicas. Los lanzadores activos están deshabilitados. La extensión de captura registra metadatos únicamente desde fixtures locales. FO1 HTTP 400 permanece sin resolver. Las capturas privadas, las credenciales, el lib...

HOL Guard AI Agent Runtime Security Benchmark

HOL Guard Benchmark is an open source fixture-based benchmark for evaluating security controls applied to AI coding agents. It defines deterministic scenarios covering secret access, risky shell execution, tool poisoning, approval handling, safe operations, and receipt generation across Codex CLI...

PECR: A Reproducible Specification and Synthetic Stress Test of Telemetry-Informed Vulnerability Prioritization for SD-WAN

Software-defined wide-area networking concentrates operational authority in controllers, orchestrators, and Internet-facing edges, but severity-only remediation queues do not represent current exposure or evidence quality. This paper specifies Predictive Exposure and Cryptographic Readiness PECR,...

The vulnerability of the Python relenv tool for creating reproducible and portable environments is related to insufficient resource control during its existence. This allows a malicious actor to trigger a service failure.

The vulnerability of the Python relenv tool for creating reproducible and portable environments is related to insufficient resource control during its existence. Exploiting this vulnerability could allow a malicious actor to cause service failures...

Rethinking AI Cloud Infrastructure for Agentic Serving Systems with the Aries Experimentation Framework

Autonomous agents challenge conventional LLM serving by coupling repeated inference with persistent context and sandboxed tool execution. We present Aries, a full-stack experimentation framework that separates task semantics from execution configurations, reconstructs cross-component agent...

Root-the-Box

EH-Stage6: Exploit Chain as Code This repository contains my...

vuln-verify

vuln-verify A vulnerability verification engine driven by evi...

pentest-confirm-mcp

pentest-confirm-mcp Model Context Protocol MCP server for...

llm-redteam-helixpay

Red-teaming an LLM support agent — HelixPay A self-contained...

GHSA-8WR5-JM2H-8R4F vLLM has Remote DoS via Invalid Recovered Token Reinjection

Summary A frontend-legal multi-request speculative workload can make vLLM produce an out-of-vocabulary recovered token equal to vocabsize, convert that value to -1 when choosing the next live token for a request, and then feed that -1 back into the next drafter input ids. On Qwen3 GPTQ this reach...

Flounder 0.2.0

Flounder turns modern coding agents into an end-to-end security audit system. Give it an authorized target boundary - a repository, source tree, package, deployed clue, or prior run - and the agent can prepare the workspace, read the code and supporting material, map the attack surface, dig into...

GitHub 'Verified' Commits Can Be Rewritten Into New Hashes Without Breaking Signatures

New research shows that a signed Git commit's hash is not the one-of-a-kind name that much of the software world assumes it to be. Given any signed commit, someone without the signing key can mint a second commit with the same files, author, and date, and a valid signature, GitHub still stamps...

[SECURITY] Fedora 44 Update: helm-4.2.2-1.fc44

Helm is a tool for managing Charts. Charts are packages of pre-configured Kubernetes resources. Use Helm to: - Find and use popular software packaged as Helm Charts to run in Kubernetes - Share your own applications as Helm Charts - Create reproducible builds of your Kubernetes applications -...

Git Hash Chain Malleability

This whitepaper demonstrates that signed Git commits are not uniquely bound to their commit hashes, even when they display a valid "Verified" badge on GitHub. The authors show that attackers can transform the signature encoding of an existing signed commit—without possessing the private signing...

Flounder 0.1.2

Flounder turns modern coding agents into an end-to-end security audit system. Give it an authorized target boundary - a repository, source tree, package, deployed clue, or prior run - and the agent can prepare the workspace, read the code and supporting material, map the attack surface, dig into...

PT-2026-53227

Name of the Vulnerable Software and Affected Versions WsgiDAV versions prior to 4.3.5 Description The sample MySQLBrowserProvider in wsgidav/samples/mysql dav provider.py concatenates the record key parsed from a request URL directly into SQL WHERE clauses. This occurs within the exists record by...