plugin-security

Professional Agent Skill for building, auditing, testing, and releasing modern WordPress plugins with Codex, Cursor, and Claude Code.

A WordPress plugin designed for educational purposes, demonstrating techniques used in plugin-based persistence and evasion, to aid in security research and training.​

Always-on compatibility testing for DeepSeek Harness plugins: exact releases, isolated runners, and fixable upstream issues.

Static, no-execution security checks for Hermes Agent plugins

WP Hunter is a bilingual command-line toolkit for collecting WordPress.org and Patchstack VDP targets, downloading verified plugin archives, and running fail-safe Semgrep triage.

Security vetting for DeepSeek Harness (DSH) plugins: permission & supply-chain audits before install, graded via the open dsh-vet/v1 report standard.

This tool scans WordPress sites for vulnerabilities in the "RegistrationMagic" plugin (CVE-2024-10508). It checks for the presence of a specific version (`6.0.2.6`) and marks the site as vulnerable if found. The results are saved in a file (`vuln.txt`) for further analysis.

Security-hardened ETS2LA fork for ETS2/ATS with verified plugin installation, fail-safe control output, reproducible builds, and Chinese developer docs.

Security-focused OBS Studio launcher with plugin integrity validation, firewall rule management, startup verification, and session protection features.

🐤 Native security canary. Probes what an untrusted plugin can access.

🛡️ WASM security canary. Proves the Wasmtime sandbox blocks every attack the native canary exposes.

Read-only DSH plugin security auditor with current-session model review

Hide commands from tab-complete, block their execution, and keep your plugin list private — per LuckPerms group.

Rust sandbox for AI agents and untrusted code, inspired by OpenAI Codex. OS-enforced isolation on Linux, Windows, and macOS. (Beta)

Reproducible dataset of abandoned WordPress plugins (10K+ active installs, 2+ years no update). Companion to royalplugins.com/plugin-graveyard/. Snapshot June 2026: 176 plugins, 4.93M installs at risk.

A plugin-native security scanner for DeepSeek Harness that detects capability escalation, config hijacking, prompt injection, dangerous code, and self-modifying behavior.

HOL Guard template: pre-configured secure plugin starter with CI

Security research, datasets, and reproducible findings from AgentGuard across AI agents, plugins, MCP ecosystems, and agent runtimes.

面向 DeepSeek Harness 的 agent 专用插件调查与安装工具桥。它不提供设置页、可视化市场、客户端 bundle 或 Web 路由。

HOL Guard examples and guides for AI agent security