Resources

Resources

📚 Awesome Lists 🤖 Securing AI 🛰️ AI for Defense ⚔️ Offensive AI 📋 Frameworks 🐳 Docker Images 🖥️ Self Hosting 📖 Books 📡 Blogs & Podcasts 🛠️ Tools
--- ## 📚 Awesome Lists

🔐 Security — General

  • Awesome Cyber Security — A collection of awesome software, libraries, documents, books, and resources about security.
  • Awesome Cyber Security (okhosting) — A curated list of cyber security resources and tools.
  • Awesome Security — A collection of awesome software, libraries, documents, books, resources and cool stuff about security.
  • Awesome Appsec — A curated list of resources for learning about application security: books, websites, blog posts, and self-assessment quizzes.
  • Awesome Security Hardening — A collection of awesome security hardening guides, best practices, checklists, benchmarks, tools and other resources.
  • Awesome Infosec — A curated list of awesome information security resources.

🎯 Offensive Security & Pentesting

🔵 Detection, DFIR & Threat Intel

🌐 Networking & Traffic Analysis

  • Awesome Network Analysis — An awesome list of resources to construct, analyze and visualize network data.
  • Awesome Networking — A curated list of awesome networking libraries, resources and shiny things.
  • Awesome PCAP Tools — List of tools to help process pcap files for network traffic research.

💻 Scripting & Development

  • Awesome Bash — A curated list of delightful Bash scripts and resources.
  • Awesome Powershell — A curated list of delightful PowerShell packages and resources.
  • Awesome Python — An opinionated list of awesome Python frameworks, libraries, software and resources.
  • Awesome README — A curated list of awesome READMEs.

🎤 Talks & Learning

--- ## 🤖 Securing AI

🔴 AI Red Teaming

Open source tools and frameworks for adversarial testing, attacking, and evaluating AI/ML systems.

  • PyRIT — Microsoft's Python Risk Identification Toolkit for generative AI. Automates red teaming of LLMs, multimodal models, and AI pipelines.
  • Garak — LLM vulnerability scanner. Tests for prompt injection, jailbreaks, data leakage, hallucination, and more across many model providers.
  • ps-fuzz (Prompt Fuzzer) — Open source tool for fuzzing LLM system prompts to find injection vulnerabilities and weaknesses.
  • promptfoo — Test and red-team LLM prompts and apps with repeatable, version-controlled eval configs.
  • AI Exploits (ProtectAI) — Real-world exploits and vulnerabilities found in ML infrastructure and AI tooling. Maintained by the ProtectAI team.
  • Adversarial Robustness Toolbox (ART) — IBM's library for defending ML models against adversarial attacks: evasion, poisoning, extraction, and inference attacks.
  • CleverHans — A Python library to benchmark ML model vulnerability to adversarial examples.
  • TextAttack — A framework for adversarial attacks, data augmentation, and adversarial training in NLP.
  • PromptBench — Microsoft's unified library for evaluating and benchmarking LLM robustness against adversarial prompts.
  • Counterfit — Microsoft's automation layer for assessing the security of ML systems across multiple attack types.
  • FuzzyAI — CyberArk's automated LLM fuzzer for surfacing jailbreaks and injection weaknesses in model APIs.
  • promptmap — A prompt-injection scanner aimed at custom LLM applications.
  • DeepTeam — An LLM red-teaming framework with 40+ vulnerability types and 10+ attack methods, mapped to the OWASP LLM Top 10.
  • GPTFuzz — Automated jailbreak-template generation for stress-testing LLM guardrails.
  • PromptInject — A framework for assembling and quantifying adversarial prompt-injection attacks.
  • PALLMs — A collected set of payloads for attacking LLMs, handy for building test suites.
  • Foolbox — A toolbox for crafting and evaluating adversarial examples against models in PyTorch, TensorFlow, and JAX.

🔵 AI Blue Teaming & Defense

Open source tools for monitoring, protecting, and securing AI/ML systems and applications in production.

  • Rebuff — Self-hardening prompt injection detector. Uses LLM-based detection, a canary token system, and a shared database of attack signatures.
  • LLM Guard — A comprehensive security toolkit for LLM interactions: input/output scanning for prompt injection, PII, toxicity, and code detection.
  • NeMo Guardrails — NVIDIA's open source toolkit for adding programmable guardrails to LLM-based conversational applications.
  • LangKit — An open source text metrics toolkit for monitoring LLM inputs/outputs. Detects prompt injection, toxicity, PII, and sentiment drift.
  • Vigil — LLM prompt injection and jailbreak detection server with REST API. Scans prompts against known attack signatures and semantic embeddings.
  • MLflow — Open source platform for ML lifecycle management including experiment tracking, model registry, and monitoring for drift.
  • Evidently AI — Open source ML and LLM observability framework. Monitors model performance, data drift, and output quality over time.
  • Guardrails AI — Open source framework for adding structural, type, and quality constraints to LLM outputs.
  • Purple Llama — Meta's suite of LLM safety tools, including Llama Guard, Prompt Guard, Code Shield, and CyberSecEval.
  • LlamaFirewall — A framework for detecting and mitigating security risks across multi-step agentic operations.
  • NB Defense — Scans Jupyter notebooks for secrets, PII, license issues, and vulnerabilities — a common blind spot in ML workflows.
  • Privacy Meter — Audits models for data-privacy risk by quantifying membership-inference exposure.
  • lintML — A security linter for ML codebases, from NVIDIA.
  • Prompt Injection Defenses — A curated catalog of practical defenses against prompt injection, maintained by tl;dr sec.

📖 AI Security Learning Resources

Open source courses, guides, research, and references for learning AI/ML security concepts.

  • OWASP Top 10 for LLMs — The OWASP project documenting the top 10 most critical vulnerabilities in LLM applications, with mitigations.
  • Prompt Engineering Guide — Comprehensive guide to prompt engineering including adversarial prompting, prompt injection, and jailbreaking techniques.
  • MITRE ATLAS — Adversarial Threat Landscape for Artificial-Intelligence Systems. A knowledge base of adversarial ML tactics and techniques analogous to ATT&CK.
  • Awesome LLM Security — A curated list of LLM security resources: papers, tools, and real-world vulnerabilities specific to language models.
  • Awesome ML Privacy — Resources on privacy in machine learning: membership inference, model inversion, differential privacy, and federated learning.
  • AI Red Team Exercises — Practical exercises and lab scenarios for hands-on adversarial AI testing and red team practice.
  • NIST AI RMF — The framework to anchor an AI security and governance program around.

🏛️ Governance, Standards & Frameworks

Standards, control frameworks, and risk taxonomies for building and auditing an AI security program.

🔌 MCP & Agent Security

Tooling and references for securing Model Context Protocol servers, connectors, and agentic workflows.

  • Awesome MCP Security — A dedicated running list of everything MCP security — attacks, defenses, and tooling.
  • MCP-Scan — A security scanner for MCP servers that flags tool poisoning, prompt injection, and risky configurations.
  • mcp-context-protector — Trail of Bits' security wrapper for MCP servers, guarding against line-jumping and config-change attacks.
  • MCP Guardian — Real-time control and mediation of an assistant's access to MCP servers.
  • Secure MCP Gateway — An MCP gateway with authentication, tool discovery, and guardrail enforcement.
  • MCP Security Checklist — SlowMist's checklist for hardening MCP-based tools and plugin ecosystems.
  • Agentic Radar — A CLI scanner that maps and analyzes agentic workflows for security weaknesses.
  • Agent Threat Rules (ATR) — A "Sigma for agents": open, machine-readable detection rules for prompt injection, tool poisoning, and MCP attacks that convert to Splunk, Elastic, and SARIF.
  • agent-audit — A forensic auditor for local coding agents (Claude Code, Codex) and a project scanner for repos shipping skills, plugins, and MCP manifests.
  • aguara — A single-binary static scanner for AI-agent skills and MCP servers, with pattern, taint-tracking, and rug-pull detection.
  • Snyk agent-scan — A scanner for AI agents, MCP servers, and skills — the maintained successor to the original mcp-scan work.
  • Cisco AI Defense mcp-scanner — Scans MCP servers and tool surfaces for tool/prompt/resource risks, package risk, and malware indicators.
  • Cisco AI Defense skill-scanner — Scans agent skills with YARA patterns, LLM-as-a-judge, and dataflow analysis.

🔍 Model & Supply-Chain Scanning

Scanners for model artifacts, serialized files, and AI bills of materials.

  • ModelScan — Protect AI's scanner for detecting unsafe code embedded in serialized model files.
  • picklescan — Detects Python pickle files that perform suspicious actions on load.
  • fickling — Trail of Bits' pickle decompiler and static analyzer for inspecting untrusted model files.
  • OWASP AIBOM — The AI Bill of Materials project for inventorying models, data, and dependencies.
  • Safetensors — A safe serialization format that avoids the code-execution risks of pickle-based model files.
  • Model Transparency — Sigstore's project for signing and verifying model provenance and integrity.
  • AIShield Watchtower — An open tool for studying AI models and scanning them for vulnerabilities.

🧪 Labs, CTFs & Practice

Intentionally vulnerable apps and challenges for hands-on AI security practice.

  • AI Goat — A deliberately vulnerable LLM app with challenges mapped to the OWASP LLM Top 10.
  • Damn Vulnerable MCP Server — An intentionally insecure MCP server for learning to attack and defend the protocol.
  • Damn Vulnerable LLM Agent — A vulnerable agent for practicing prompt injection, tool misuse, and agent attacks.
  • AI Red Teaming Playground Labs — Microsoft's runnable labs and infrastructure for AI red-team training.
  • Gandalf — Lakera's browser-based prompt injection game — the gentlest possible intro to the problem.
  • Prompt Hacking Resources — A curated collection for learning AI red teaming, jailbreaking, and prompt injection.
  • Learn Prompting — Prompt Hacking — Free structured courses on prompt hacking and AI safety, plus the HackAPrompt competition.

📡 Communities, Podcasts & Lists

Places to keep current, plus the curated lists these resources were drawn from.

--- ## 🛰️ AI for Defense

🧭 Threat Intel & Detection

Using LLMs to accelerate CTI, IOC/TTP extraction, and detection content.

  • trs — Summarizes threat reports and extracts MITRE TTPs and IOCs with an LLM and vector store.
  • TI-Mindmap-GPT — Turns threat intel into AI summaries, mindmaps, IOC/TTP extraction, and ATT&CK Navigator layers.
  • ThreatIngestor — Extracts and aggregates IOCs from feeds and integrates with MISP — a natural pairing with LLM post-processing.
  • AttackGen — Generates tailored incident-response scenarios from MITRE ATT&CK and ATLAS.

🛎️ SOC & Alert Triage

LLM assistants for alert triage, investigation, and SOC automation.

  • SOCGPT — LLM-driven log summarization, severity triage, and MITRE mapping for SOC work.
  • AI-SOC-Agent — A Black Hat 2025 MCP server exposing security-investigation tools (ELK, IRIS) to an agent.
  • agentic-soc-platform — An agentic SOC platform on LangGraph/Dify with local-LLM support.

🔎 Code Review & DFIR

LLM-assisted vulnerability discovery, code review, and forensic triage.

  • Vulnhuntr — Zero-shot vulnerability discovery in Python repos via LLM call-chain analysis; credited with a real 0-day RCE.
  • Claude Code Security Review — An official Claude-based semantic SAST GitHub Action that reviews PR diffs for vulnerabilities.
  • MemoryInvestigator — Volatility 3 + LLM + RAG for memory-forensic triage.
  • Volatility-MCP-Server — An MCP server exposing Volatility 3 plugins for natural-language memory forensics.
  • EscalateGPT — Uses GPT to find privilege-escalation paths in AWS IAM policies.

🍯 LLM Honeypots & Deception

Honeypots that use LLMs to simulate convincing systems — a natural homelab fit.

  • Beelzebub — A low-code honeypot that uses LLMs to simulate SSH, HTTP, and even MCP services.
  • shelLM — An LLM-powered SSH honeypot from the Stratosphere lab ("LLM in the Shell").
  • VelLMes — A multi-protocol LLM honeypot framework, successor to shelLM.
--- ## ⚔️ Offensive AI

🗡️ Pentest & Red-Team Agents

Autonomous and semi-autonomous AI agents for penetration testing — for authorized testing only.

  • PentestGPT — The original USENIX'24 LLM pentest agent, now a fuller autonomous pipeline.
  • PentAGI — A fully autonomous multi-agent pentest framework with Docker sandboxing.
  • CAI (Cybersecurity AI) — A modular, bug-bounty-ready agent framework supporting 300+ models.
  • Strix — Autonomous "AI hackers" that run code dynamically and validate findings with PoCs.
  • hackingBuddyGPT — A minimal research framework for LLM-driven privilege escalation and web pentesting.
  • HexStrike-AI — An MCP server exposing 150+ security tools (nmap, nuclei, gobuster, …) to AI agents.

🎯 CTF, Exploit & Bug-Bounty Benchmarks

Offensive agents and the benchmarks used to measure their capability.

  • SWE-agent (EnIGMA) — SWE-agent's offensive-CTF mode, state-of-the-art on NYU CTF and Cybench.
  • Cybench — 40 professional CTF tasks, widely used by AI safety institutes to measure cyber capability.
  • NYU CTF Bench — Dockerized CSAW CTF challenges for evaluating LLM agents.
  • BountyBench — 25 real systems and 40 bug bounties for detect-exploit-patch evaluation.

🧬 LLM-Driven Fuzzing

Using LLMs to generate fuzzing harnesses and targets for traditional software.

  • oss-fuzz-gen — Google's LLM-driven fuzz-harness generation for OSS-Fuzz; found real CVEs, including one in OpenSSL.
  • Fuzz4All — A "universal" LLM-based fuzzer across compilers and languages (ICSE 2024).
  • ChatAFL — LLM-guided protocol fuzzing that extends AFLNet (NDSS'24).

🔬 AI-Assisted Reverse Engineering

LLM plugins for binary analysis and traffic inspection.

  • Gepetto — An IDA Pro plugin that uses GPT to add comments and rename variables.
  • GhidraMCP — An MCP server exposing Ghidra's reverse-engineering operations to any MCP-capable LLM.
  • burpgpt — A Burp Suite extension integrating GPT for additional passive scanning.
--- ## 📋 Frameworks & Standards

🗺️ Adversary & Defense Models

  • MITRE ATT&CK — The knowledge base of adversary tactics and techniques that most modern detection work maps back to.
  • MITRE D3FEND — ATT&CK's defensive counterpart — a knowledge graph of countermeasures mapped to attacker techniques.

📐 Controls, Standards & Compliance

  • NIST Cybersecurity Framework — The high-level Identify / Protect / Detect / Respond / Recover framework for organizing a security program.
  • NIST SP 800-53 — The comprehensive catalog of security and privacy controls that underpins most control mappings.
  • CIS Controls — A prioritized, prescriptive set of safeguards — a practical starting point when NIST feels too broad.
  • CIS Benchmarks — Consensus hardening baselines for operating systems, cloud platforms, and applications.
  • OWASP — The home of the Top 10, ASVS, cheat sheets, and much of the appsec canon.
  • PCI SSC (PCI DSS) — The official source for the Payment Card Industry Data Security Standard and its supporting documents.
--- ## 🐳 Docker Images for Security

🗡️ Offensive & Pentesting

🎯 Vulnerable Training Targets

🔬 Scanning & Assessment

--- ## 🖥️ Self Hosting
  • Awesome Selfhosted — A list of Free Software network services and web applications which can be hosted on your own servers.
--- ## 📖 Books

🔵 Detection, DFIR & Blue Team

  • The Practice of Network Security Monitoring — Richard Bejtlich. The foundational case for NSM and why detection has to complement prevention.
  • Applied Network Security Monitoring — Chris Sanders & Jason Smith. A hands-on walk through collection, detection, and analysis.
  • Practical Packet Analysis — Chris Sanders. The go-to for getting genuinely comfortable in Wireshark.
  • Intelligence-Driven Incident Response — Scott J. Roberts & Rebekah Brown. Ties threat intel and IR together around the F3EAD loop.
  • Crafting the InfoSec Playbook — Jeff Bollinger, Brandon Enright & Matthew Valites. Building detection and response as a repeatable program.
  • Blue Team Handbook: Incident Response Edition — Don Murdoch. A dense field reference for responders.

🤖 AI Security

  • Not with a Bug, But with a Sticker — Ram Shankar Siva Kumar & Hyrum Anderson. An accessible tour of how ML systems fail and get attacked.
  • Adversarial AI Attacks, Mitigations, and Defense Strategies — John Sotiropoulos. A practitioner's guide to securing ML and LLM systems.

📚 Narrative & Big Picture

  • Sandworm — Andy Greenberg. The story of the most consequential nation-state attacks of the last decade.
  • Countdown to Zero Day — Kim Zetter. The definitive account of Stuxnet.
--- ## 📡 Blogs, Newsletters & Podcasts

📰 Newsletters & Blogs

  • tl;dr sec — Clint Gibler's weekly roundup of the best security tools, talks, and research. One of the most-read in the field.
  • Detection Engineering Weekly — Zack Allen's curated digest of detection engineering content, tooling, and threat news.
  • SANS Internet Storm Center — Daily handler diaries on emerging threats and traffic anomalies.
  • Black Hills Information Security — Practical offensive and defensive posts, webcasts, and tooling from the BHIS crew.
  • The DFIR Report — Deep, artifact-level walkthroughs of real intrusions — excellent detection-writing fodder.

🎧 Podcasts

  • Risky Business — Weekly news and analysis with a sharp practitioner's take. Also a great newsletter.
  • Darknet Diaries — Narrative storytelling from the dark side of the internet — the one to recommend to non-specialists.
--- ## 🛠️ Tools

🚨 Detection Engineering

  • Sigma — The generic, vendor-agnostic detection rule format and its community rule repo. The lingua franca if you write detections anywhere.
  • Uncoder.io — Translates detection rules between SIEM dialects (Sigma → Splunk, Sentinel, Elastic, and dozens more). Handy when you're moving logic across platforms.
  • detection.fyi — A searchable front-end for the Sigma rule set — faster than grepping the repo when you want to see how others detect a technique.
  • Elastic Detection Rules — Elastic's open detection rule repo. Worth reading even if you're not on Elastic; the logic translates.
  • Splunk Security Content — Splunk's ESCU analytics as code — another strong library to learn detection patterns from.
  • ATT&CK Navigator — Layer your detection coverage over the ATT&CK matrix to see where the gaps are.
  • DeTT&CT — Scores your data sources and detection coverage so you can prioritize what to build next.

🪵 Log Parsing & Analysis

  • CyberChef — The "cyber swiss army knife" — decode, deobfuscate, and transform data right in the browser. You'll reach for it constantly.
  • Grok Debugger — Kibana's built-in tool for building and testing grok patterns against real log samples before they hit the pipeline.
  • regex101 — Live regex tester with an explanation pane — indispensable when a pattern won't match and you can't see why.

🕵️ DFIR & Threat Hunting

  • Velociraptor — Endpoint visibility and DFIR collection at scale — VQL-driven hunting across the fleet. My write-up: Security Lab Tool Install Part 2.
  • OpenCTI — Open-source threat intelligence platform for storing and correlating CTI. My write-ups: Lab Install · Wazuh → OpenCTI Integration.
  • MISP — The classic threat-intel sharing platform for IOCs and indicators.
  • Chainsaw — Fast Sigma-based hunting through Windows Event Logs and other forensic artifacts.
  • Hayabusa — Sigma-based Windows event log timeline generator — great for first-response triage.
  • Wireshark — The packet analyzer. Still the one everyone opens.
  • Zeek — Turns network traffic into rich, structured logs for hunting and monitoring.
  • Eric Zimmerman's Tools — The essential Windows forensics toolkit — MFT, registry, shellbags, and more.

📇 Reference & Lookup

  • GTFOBins — How trusted Unix binaries get abused for privesc and living-off-the-land. Pairs with the next one.
  • LOLBAS — The Windows equivalent — living-off-the-land binaries, scripts, and libraries.
  • Exploit-DB — Searchable archive of public exploits and PoCs.
  • LetsDefend — Hands-on blue-team practice — investigate real alerts in a simulated SOC.
  • Blue Team Labs Online — Defensive CTF-style labs and investigations to sharpen detection and IR skills.