SEW

SEW: Style-Encoded Watermarking of LLM-Generated Code Paper Link 📖 📰 News 📢 NEW! The SEW paper is available on arXiv: arXiv:2609.39414. Sep 30, 2026 📢 The official SEW code has been released on GitHub. Sep 30, 2026 🔍 Motivation Feature| Token-level watermarks KGW, SWEET, Unigram, STONE, STA-1|...

watermarks-remover

| | | || | | |/ |/| || |/ |/ |/ | |/| | | | | | |/ ||| | | | | | \ | | | | | \ | \ | \ | | | || / | | \ watermarks-remover Agent skill + stdlib Python service to strip multi-vendor AI provenance marks from text and files — for privacy and hygiene on content you own. The skill is a thin client:...

OWASP-Model-Card-Security-Standard

Status: Draft — Incubating OWASP Project This standard is under active development. Contributions and feedback are very welcome. What is this? When you download a pre-trained model from HuggingFace or any other repository, you're essentially trusting a black box. You don't know what data it was...

aegisagent

AegisAgent — AI Agent Runtime Security Gateway AI Agent Runtime Security Gateway — cryptographically binds every tool call to the authentic model completion, zero-trust policy enforcement, and integrated sandbox isolation and human approval. Background: Why AegisAgent In August 2026, the Cloud...

linux-kernel-codex-harness-v2

Kernel Codex Harness v2 한국어 | English Research Tool · Original Import: 3 April 2026 · v2 Documentation Revision: 11 July 2026 External Signal: From Attention Allocation to Provenance-Aware Triage Use reproducible observations to guide model attention, then use repository provenance to organize...

trustlock

trustlock A Git-native dependency admission controller. Evaluates trust signals on every dependency change. How it works trustlock runs as a Git pre-commit hook advisory mode and a CI check enforce mode: Advisory pre-commit: warns on violations, exits 0, advances the trusted baseline when all...

agensic

Know what ran, where, when and who executed it The missing observability layer for AI-powered terminal workflows Signed Provenance Tracked Sessions Git Time Travel Replayable Sessions IDE style Tab Autocomplete Agensic upgrades your existing terminal workflow for the AI era. It is built for...

demo-cve-2026-4821

Demo: CVE-2026-4821 Vulnerability Remediation This repository demonstrates how OpenExecution's provenance system tracks AI agent CVE remediation with full cryptographic auditability. Scenario Vulnerability : SQL injection in the authentication module Agent : sentinel-x9 CyberSafe Inc. Flow :...

ZoneClaw-code

ZoneClaw Experiment code for ZoneClaw: Mitigating Persistent Memory Attack by Establishing Memory-Zoning among OpenClaw-Style Computer-Use Agents. This repository contains the benchmark tasks, defense implementations, and evaluation scripts used in the paper. The benchmark studies cross-session...

nullorigin

NullOrigin 🛡️ Universal AI Provenance & Watermark Sanitization Middleware Research artifact — for watermarking robustness evaluation only. 🔬 Research use only NullOrigin is a research artifact. It is published to support academic and independent study of watermarking robustness, and for no other...

SkillLifeBench: Benchmarking Security across the Lifecycle of Agent Skills

Web agents reuse skills whose stored code, selected version, and loaded implementation can diverge across tasks and updates. SkillLifeBench links these states to task outcomes in 388 paired scenarios spanning implementation, interface, runtime, composition, and version changes. We evaluate 14...

SoK: Timestamps Are Not a Time Model: Rethinking Temporal Reasoning in Host-Based Intrusion Detection

Host-based intrusion detectors order events, window audit streams and sequence attack stages by their timestamps. Every conclusion they draw therefore rests on assumptions about clocks, delivery, entity identity and the integrity of the evidence, and most of these assumptions go unstated. This...

Back to Basics: Formal Analysis Hints Improve Weak-To-Strong Monitoring

Frontier agentic systems are capable of taking misaligned actions that may cause significant harm. Monitoring techniques are developed to oversee these systems in order to detect and prevent dangerous actions. However, current state-of-the-art monitors suffer from four persistent failure modes:...

AgentTriage: Replayable Security Diagnosis across Heterogeneous Agent Runtimes

When tool-use agents write files, execute code, or send data, their final responses do not explain the resulting security events, and heterogeneous runtimes expose the relevant evidence through incompatible interfaces. We present AgentTriage, an executable contract for replayable security diagnos...

Agentic AI Defense: Decode-Time Provenance Binding for MCP

Agentic AI systems using the Model Context Protocol MCP are vulnerable to indirect prompt injection through untrusted tool outputs. We demonstrate cross-modal trust laundering, in which instructions concealed in an email-signature image using least-significant-bit LSB, discrete cosine transform...

Compaction Laundering: Provenance Destruction in Context-Compacting Coding Agents

Coding agents that run long enough must compact their context, summarizing the conversation, discarding the original, and carrying the summary forward as memory. We show that in popular open-source coding agents this step is an exploitable trust boundary. Harnesses that flatten the typed...

SUSE-SU-2026:24282-1 Security update for helm

This update for helm fixes the following issues: - CVE-2026-33814: golang.org/x/net/http2: infinite loop in HTTP/2 transport when given bad SETTINGSMAXFRAMESIZE bsc1265758. - CVE-2026-35204: github.com/helm/helm: helm.sh/helm/v4: Helm: Arbitrary file write via specially crafted plugin bsc1261939....

PROVISO: Provenance-Informed Action Selection for Secure LLM Agents

LLM agents act on the external world through tools and third-party skills, so a single executed action can cause irreversible harm. Existing defenses judge a proposed action without tracking whether its content comes from the user or from external sources such as tool outputs, and they judge it i...

What It Is, Not What It Says: Securing Agentic Browsing Interactions with Web Semantics

AI agents increasingly browse the web on users' behalf, yet they consume webpages as screenshots or text, exposing their LLMs to unvetted third-party content. Browsers enforce semantics under which what an element can do depends on what it is and where it comes from, not on what it says. Agents...

kin

Una nueva base para el código. Kin es un repositorio de código nativo de grafos para personas y agentes de IA. Almacena el código fuente, las relaciones de código registradas y el historial versionado como estado del repositorio. El grafo es el modelo del repositorio, no un índice de búsqueda...