claude-code-guardrails

We have since built a dedicated product to help developers manage AI permissions better, which we intend to make open source. Check it out atstoplight.ai. Claude Code Guardrails settings.json works if: You're fine editing JSON and killing Claude Code sessions every time policy changes Your securi...

tealtiger

TealTiger AI Agent Security & Governance SDK Deterministic governance, guardrails, cost tracking, and policy management for LLM applications. Open source. TypeScript + Python. Works with any provider. Website · Documentation · Examples · Discord · Contributing ⚡ 60-second quickstart Install: npm...

Guardrails

NVIDIA NeMo Guardrails Library LATEST RELEASE / DEVELOPMENT VERSION : The develop branch tracks the latest top of tree development. The latest released version is 0.24.1. ✨✨✨ 📌 The official NeMo Guardrails library documentation is available atdocs.nvidia.com/nemo/guardrails. ✨✨✨ NVIDIA NeMo...

LiteLLM < 1.25.0 - Remote Code Execution

LiteLLM before 1.25.0 allows authenticated users with the master API key to execute arbitrary Python code through the /guardrails/testcustomcode endpoint intended for custom guardrail testing. The endpoint’s insufficient filtering mechanisms can be bypassed, allowing attackers to abuse Python’s...

CKA-Agent

ICML 2026 & ICLR 2026 AIWILD CKA-Agent: Bypassing LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search 🛡️ Defense towards CKA TurnGate is a response-aware defense mechanism designed to detect and mitigate hidden malicious intent in multi-turn dialogue systems. Defending...

CorrectGuard: Eyes-Off Correctness Estimation for Black-Box Security Guardrails

AI services increasingly rely on black-box security guardrails, yet privacy-preserving model auditing regimes often cannot measure how well these systems perform in both a human eyes-off production setting, which disallows human inspection of user input, and a machine eyes-off setting, which...

PT-2026-105791

Summary Decepticon wraps web crawl results — the output of agent reconnaissance against target services — into LLM messages without neutralizing ChatML special-token literals. Under the BYOK Bring Your Own Key deployment model, users configure their own LLM credentials to any OpenAI-compatible...

Guardrails v0.24.1

NVIDIA NeMo Guardrails Library LATEST RELEASE / DEVELOPMENT VERSION : The develop branch tracks the latest top of tree development. The latest released version is 0.24.1. ✨✨✨ 📌 The official NeMo Guardrails library documentation is available atdocs.nvidia.com/nemo/guardrails. ✨✨✨ NVIDIA NeMo...

CISO's Expert Guide to Agentic Pentesting for Websites

Attackers now weaponize new vulnerabilities in about five days Mandiant, part of Google Cloud. The median organization takes 43 days to patch one Verizon DBIR 2026. A new free guide explains how autonomous AI agents are closing that gap, and what security leaders must demand before pointing one a...

CS-Guard: Benchmarking LLM Guardrails for Code Generation Security

Large language models LLMs have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains unclear. We introduce CS-Guard, the first benchmark to systematically evalu- ate guardrails for code generation security. It covers 1 text-to-code...

Architecting the Secure AI-SOC: A Neurosymbolic Framework for Pipeline Integrity and Threat Mitigation

The integration of Large Language Models LLMs into Security Operations Centers SOCs streamlines threat intelligence but introduces critical vulnerabilities, notably indirect prompt injection via log poisoning. Adversaries exploit this vector to execute multistep "promptware'' kill chains by...

CVE-2026-82855 @hulumi/policies before 1.3.2 Evidence Validation Bypass

@hulumi/policies versions before 1.3.2 contain an evidence validation bypass vulnerability in Cloudflare and deployment-governance validators that allows attackers to suppress violations by submitting unrelated compliant evidence. Attackers can use evidence from different zones, hostnames, origin...

EUVD-2026-68443

@hulumi/policies versions before 1.3.2 contain an evidence validation bypass vulnerability in Cloudflare and deployment-governance validators that allows attackers to suppress violations by submitting unrelated compliant evidence. Attackers can use evidence from different zones, hostnames, origin...

CVE-2026-82855: Protection Mechanism Failure

@hulumi/policies versions before 1.3.2 contain an evidence validation bypass vulnerability in Cloudflare and deployment-governance validators that allows attackers to suppress violations by submitting unrelated compliant evidence. Attackers can use evidence from different zones, hostnames, origin...

“Sorry, I can’t help with that”: How your guardrails might become the attacker’s best friend

Welcome to this week's edition of the Threat Source newsletter. Hello, everyone. Long time reader, first time writer here at the Threat Source newsletter! I wanted to start out by introducing myself. My colleague and friend Mick Baccio set the bar pretty high last week, so I was planning to tell...

The Guard That Cried Wolf: How Scary Words Make Agent Guardrails Refuse Legitimate Actions

Agent guardrails are checks that approve or refuse each action before an LLM executes it. Sometimes they refuse requests that are genuinely safe. This over-safety blocks deployment when a guardrail refuses an authorized task. Evaluating over-safety is hard: at the boundary an authorized action...

Grok fooled into stealing user chat, location data, and more

A new type of prompt injection attack shows why giving AI assistants access to browsers, code tools, and private data deserves extra caution. AI researchers describe "Cryptographic Context Injection"—an attack that hides malicious instructions inside encrypted data. The AI is then persuaded to...

The safety penalty: Reclaiming operational sovereignty in the age of AI

As frontier models advance in cyber capability, their guardrails also become more restrictive. Defenders relying on these models to power core SOC processes cannot afford to pay the "safety penalty" of being blocked by these safeguards. Organizations should monitor model refusal rates and use the...

GHSA-J659-8XH6-5PQ5 atomic-agents-stack: Parallel helper/delegate batch reserves $0 for models absent from the pricing table, bypassing the cost-cap fan-out guard

estimatebatchcost atomicagents/agent.py looks up the per-model output price with PRICING.getmodel, , returning 0.0 for any model not in the hardcoded pricing table. checkbatchreservation then early-returns when the reservation is 0 and that an over-cap unknown-model batch raises...

Prompt Injections for Defense

This seems to work: Researchers from Tracebit on Monday said they found that placing prompt injections alongside passwords, cryptographic keys, and other secrets stored on Amazon Web Services was often all that was needed to shut down attacks from AI hacking agents. The prompts direct the attacki...