Tag: AI Agent Security

Cisco Skill Scanner is the first open-source security tool purpose-built for AI agent skills. Detect prompt injection, data exfiltration, and malicious code in OpenAI Codex, Cursor, and Claude Code skills using multi-engine static, behavioral, and LLM analysis.

Clawdstrike is a fail-closed policy engine and cryptographic attestation runtime for AI agent systems. Learn how this open-source EDR alternative stops shadow agents, signs every decision with Ed25519 proofs, and provides formally verified security enforcement at the tool boundary.

Crust is an open-source AI agent security gateway that intercepts and blocks dangerous tool calls before execution. With 42 built-in rules, 51 DLP patterns, and zero-config setup, it protects your secrets from prompt injection and exfiltration attacks.

DefenseClaw by Cisco AI Defense is an open-source security governance layer that scans and blocks dangerous AI agent actions before they execute. Learn installation, real code examples, and why traditional security tools fail against agentic AI threats.