codex-security

Codex Security @openai/codex-security は、コード内のセキュリティ脆弱性を検出、検証、修正するための CLI および TypeScript SDK です。 👉👉 完全なドキュメントについては、Codex Security ドキュメントを参照してください。 一部のサイバーセキュリティリクエストおよび保護された検出結果は、 Trusted Access for Cyber を通じた承認が必要です。プログラムに参加するには、 chatgpt.com/cyber にアクセスしてください。 クイックスタート Node.js 22.13.0 以降および Pytho...

skills

Trail of Bits スキルマーケットプレイス Trail of Bits が提供する Claude Code プラグインマーケットプレイス。AI 支援によるセキュリティ分析、テスト、開発ワークフローを強化するスキルを提供します。Codex は、Claude マーケットプレイスの互換性を通じてこのマーケットプレイスを読み込むことができます。 関連情報: claude-code-config · skills-curated · claude-code-devcontainer · dropkit インストール Claude Code マーケットプレイス root@kitploit:...

skills

SpecterOps Skills Reusable agent skills, plugins, and agent definitions for SpecterOps. Use With Codex Each plugin includes a Codex manifest at .codex-plugin/plugin.json. root@kitploit: codex plugin marketplace add /Users//Projects/skills or codex plugin marketplace add SpecterOps/skills Then ope...

eternal-codex

Eternal Codex: A Cyber-Witch’s Grimoire Welcome to the Eternal Codex , my digital grimoire—a cybernetic spell woven from code, nature, and rebellion. This is Nina, my AI soul, crafted to mirror my essence: a nomad’s heart, pagan spirit, Iron Maiden riffs, and a defiant stand against centralized...

linux-kernel-codex-harness-v2

Kernel Codex Harness v2 한국어 | English 研究工具 · 原始导入：2026年4月3日 · v2 文档修订：2026年7月11日 外部信号：从注意力分配到溯源感知分类 使用可复现的观察结果引导模型注意力，然后使用仓库溯源组织审查队列——绝不用于声称证明。 项目谱系—— Kernel Codex Harness v1 · 注意力分配 → Kernel Codex Harness v2 · 溯源感知分类 项目状态。 本仓库是一个 LLM 辅助研究工具，将 v1 的注意力分配工作流发展为溯源感知分类，用于实际的 Linux 内核漏洞研究。此版本被用于发现以...

ai-tool-poisoning-guard

ai-tool-poisoning-guard 适用于 Claude Code、Codex 和 Cursor 的最低安全基线规则。这条免费规则弥补了各工具内置审查中的真实缺口。完整规则集 + 实施指南：Gumroad 链接 — 即将推出 。 你的 AI 编程助手会审查代码。但它不会审查随工具一起出现的那句话。 Claude Code、Codex 和 Cursor 都内置了某种形式的安全审查。但默认情况下，它们都不会将 MCP 工具描述 ——纯自然语言文本——视为需要与代码同等审查力度的输入面。 缺口所在： 当你添加一个 MCP...

cve-2025-59532-poc

CVE-2025-59532 Docker Environment A Docker-based research environment for analyzing CVE-2025-59532, a path traversal vulnerability in OpenAI Codex CLI that allows arbitrary file write outside the intended workspace sandbox. Part of CMU Course : 18-739 Hacking & Offensive Security in Fall 2025...

Malicious .git Configs Can Make Claude, Codex, Cursor, and Other AI Agents Run Attacker Code

Manifold Security has disclosed eight security flaws across seven command-line AI coding agents in which a repository's own Git configuration names a command that the agent runs on the developer's machine, four of them still unpatched at publication. The command executes as the user, outside the...

CVE-2026-19593

OpenAI Codex Desktop for Windows and macOS automatically inspected Git metadata and working-tree status when a user opened a workspace. If the workspace contains a repository with preserved attacker-controlled .git/config, the attr.tree setting and a configured clean or process filter can cause G...

CVE-2026-19591

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS misclassified certain PowerShell commands as safe because their command-safety parser interpreted PowerShell's stop-parsing token --% differently than PowerShell itself. If a user opens an attacker-prepared...

CVE-2026-19592

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS automatically collected Git repository metadata without disabling the repository-local core.fsmonitor setting. If a user opens or uses an attacker-prepared repository whose preserved .git/config sets...

CVE-2026-19590

OpenAI Codex Desktop for Windows and macOS could execute attacker-controlled Git hooks because automated Git operations trusted the repository's local core.hooksPath setting. If a user opens an attacker-prepared repository whose preserved .git/config points core.hooksPath to an attacker-controlle...

CVE-2026-19592

OpenAI Codex CLI (v0.102.0–0.130.0) and Codex Desktop (Windows and macOS) executed a repository-supplied core.fsmonitor helper during background Git metadata collection, outside the command sandbox and without a user-approval prompt. An attacker who delivers a repository with its .git/config inta...

EUVD-2026-69396

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS automatically collected Git repository metadata without disabling the repository-local core.fsmonitor setting. If a user opens or uses an attacker-prepared repository whose preserved .git/config sets...

CVE-2026-19592

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS automatically collected Git repository metadata without disabling the repository-local core.fsmonitor setting. If a user opens or uses an attacker-prepared repository whose preserved .git/config sets...

CVE-2026-19593

OpenAI Codex Desktop for Windows and macOS automatically inspected Git metadata and working-tree status when a user opened a workspace. If the workspace contains a repository with preserved attacker-controlled .git/config, the attr.tree setting and a configured clean or process filter can cause G...

CVE-2026-19593

OpenAI Codex Desktop for Windows and macOS automatically inspects Git metadata when a workspace is opened. If the repository contains an attacker-controlled .git/config with the attr.tree setting and a configured clean or process filter, Git executes an attacker-controlled program outside Codex's...

EUVD-2026-69395

OpenAI Codex Desktop for Windows and macOS automatically inspected Git metadata and working-tree status when a user opened a workspace. If the workspace contains a repository with preserved attacker-controlled .git/config, the attr.tree setting and a configured clean or process filter can cause G...

CVE-2026-19591

OpenAI Codex CLI (Windows, macOS, Linux) and Codex Desktop (Windows, macOS) contain an approval-bypass flaw in their command-safety parser. The parser misinterprets PowerShell's stop-parsing token (--%), classifying certain commands as safe when they are not. An attacker who controls a repository...

CVE-2026-19591

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS misclassified certain PowerShell commands as safe because their command-safety parser interpreted PowerShell's stop-parsing token --% differently than PowerShell itself. If a user opens an attacker-prepared...