
airgap
Security for the modern age of AI: defend against bad AI agents and malicious npm packages

Security for the modern age of AI: defend against bad AI agents and malicious npm packages

Static security scanner for AI agent skill packages. Detects malicious SKILL.md files and bundled scripts before they run.

Proof-of-concept for CVE-2021-26700: remote code execution in the VSCode npm-script extension via malicious workspace settings.json, with detailed…

Static security analysis for npm packages. Detects obfuscated code, malicious patterns, and known vulnerabilities before installation.

Public testing data. Samples of log4j library versions to help log4j scanners / detectors improve their accuracy for detecting CVE-2021-45046 and…

The vibe-coding security sentinel. Apache-2.0 agentic security toolkit for AI-assisted projects: 5 deterministic scouts + LLM Brain Layer (BYOK…

Host-agnostic pre-write security hook for coding agent: detects user-input patterns via Semgrep and emits deterministic, no-LLM security guidance.

The credit score for npm packages. Analyze package reputation, maintenance, security, publisher trust, and ecosystem health before you install any…

Reproduces a stack-exhaustion denial-of-service in deepmerge-ts before 8.0.0, documents exploitation, and includes a scanner for vulnerable…

Scanner for CVE-2025-55182 (React) and CVE-2025-66478 (Next.js) - Track and remediate a critical React Server Components (RSC) / Flight protocol…

Live recon and posture auditing for AI agent infrastructure: scans MCP configs, session logs, and APIs for secrets, poisoned catalogs, and CoT leaks.

Scanner for the keyv/cacheable supply-chain attack: detects compromised npm packages, verifies payload hashes, and finds persistence implants in repo…

Detailed analysis of a critical pre-authentication out-of-bounds write vulnerability in libssh2 leading to remote code execution, with root cause,…

End-to-end simulation of a Python dependency confusion attack, sudo privilege escalation (CVE-2025-32463), and rootkit-based persistence - with full…

Proof-of-concept for Git LFS pointer poisoning, with Bash and Python scripts that create a malicious repository, simulate a rogue LFS server, and…

Exploits CVE-2026-21005 by poisoning Docker Registry V2 Schema 1 manifests via unauthenticated pushes, enabling tag overwrite and supply-chain…

Demonstrates CVE-2026-1122 Ed25519 signature bypass via low-order point injection, forging malicious IoT firmware updates with Python and C verifier…

Proof-of-concept for CVE-2026-22008 demonstrating AWS Lambda layer injection via untrusted ARNs, enabling arbitrary code execution and credential…