From f7a69cbb89ac2a8ed37ec2db516efd11cb100bce Mon Sep 17 00:00:00 2001 From: requie Date: Sun, 19 Apr 2026 11:57:12 +0000 Subject: [PATCH] Fix broken internal navigation links in README Add explicit HTML anchor IDs before each referenced ## heading to ensure reliable navigation despite emoji prefixes in heading text. Also fix the top navigation bar links which used shorthand targets that didn't match any heading in the document. https://claude.ai/code/session_019shWmATkSgUa5VeSjFTAJ7 --- README.md | 38 +++++++++++++++++++++++++++++++++++++- 1 file changed, 37 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 3e22e4a..3d0d6af 100644 --- a/README.md +++ b/README.md @@ -9,7 +9,7 @@ **A comprehensive guide to adversarial testing and security evaluation of AI systems, helping organizations identify vulnerabilities before attackers exploit them.** -[Overview](#overview) • [Frameworks](#frameworks) • [Methodologies](#methodologies) • [Tools](#tools) • [Case Studies](#case-studies) • [Resources](#resources) +[Overview](#overview) • [Frameworks](#key-frameworks-and-standards) • [Methodologies](#ai-red-teaming-methodology) • [Tools](#red-teaming-tools) • [Case Studies](#real-world-case-studies) • [Resources](#resources-and-references) @@ -42,6 +42,8 @@ --- + + ## 🎯 Overview As artificial intelligence systems become increasingly integrated into critical business operations, healthcare, finance, and decision-making processes, ensuring their security and reliability has never been more important. AI red teaming has emerged as a fundamental security practice that helps organizations identify vulnerabilities before they can be exploited in real-world scenarios. @@ -65,6 +67,8 @@ This comprehensive guide is designed for: --- + + ## 🤖 What is AI Red Teaming? **AI Red Teaming** is a structured, proactive security practice where expert teams simulate adversarial attacks on AI systems to uncover vulnerabilities and improve their security and resilience. Unlike traditional security testing that focuses on known attack vectors, AI red teaming embraces creative, open-ended exploration to discover novel failure modes and risks. @@ -94,6 +98,8 @@ AI red teaming adapts military and cybersecurity red team concepts to the unique --- + + ## 🚨 Why AI Red Teaming Matters ### The Urgency of AI Security @@ -124,6 +130,8 @@ Article 15 of the European Union AI Act obliges operators of high-risk AI system --- + + ## 📚 Key Frameworks and Standards ### NIST AI Risk Management Framework @@ -242,6 +250,8 @@ The Cloud Security Alliance's Agentic AI Red Teaming Guide explains how to test --- + + ## 🔬 AI Red Teaming Methodology ### Phase 1: Planning and Threat Modeling @@ -471,6 +481,8 @@ For each vulnerability: --- + + ## 🎯 Threat Landscape ### Adversary Types @@ -525,6 +537,8 @@ For each vulnerability: --- + + ## ⚔️ Attack Vectors and Techniques ### 1. Prompt-Based Attacks @@ -797,6 +811,8 @@ Example: Social engineering one agent to leak data to another --- + + ## 🛠️ Red Teaming Tools ### Open-Source Tools @@ -1133,6 +1149,8 @@ bun start --- + + ## 📊 Real-World Case Studies ### Case Study 1: Microsoft's SSRF Vulnerability (2024) @@ -1314,6 +1332,8 @@ Samsung employees accidentally leaked confidential company data by entering sens --- + + ## 👥 Building Your Red Team ### Team Composition @@ -1501,6 +1521,8 @@ Samsung employees accidentally leaked confidential company data by entering sens --- + + ## ✅ Best Practices ### 1. Start Early in Development @@ -1752,6 +1774,8 @@ Red team members should feel comfortable: --- + + ## 🚀 Implementation Quickstart (30/60/90) Use this phased plan to turn guidance into an operating program. @@ -1779,6 +1803,8 @@ Use this phased plan to turn guidance into an operating program. --- + + ## 🧪 Evaluation Harness (Reference Implementation) A lightweight structure for repeatable red-teaming and regression tracking: @@ -1814,6 +1840,8 @@ security-evals/ --- + + ## 🕸️ Agentic AI Attack Trees + Controls Mapping Use attack trees to connect offensive testing paths to defensive controls. @@ -1853,6 +1881,8 @@ Use attack trees to connect offensive testing paths to defensive controls. --- + + ## 📈 AI Harm Severity and Triage Model Use CVSS as a base, then add AI-specific modifiers: @@ -1873,6 +1903,8 @@ Use CVSS as a base, then add AI-specific modifiers: --- + + ## 🧩 Secure SDLC Integration Artifacts To reduce "one-off" testing, integrate red-team controls into delivery workflows. @@ -2071,6 +2103,8 @@ Starter artifacts in `templates/`: - `case-study-template.md` + + ## 📋 Regulatory Compliance ### United States @@ -2144,6 +2178,8 @@ Recommends adversarial testing before deployment and continuous monitoring in pr --- + + ## 📚 Resources and References ### Official Frameworks