Resources

Readings, tools, and references for CIS400. Assigned readings appear on the schedule next to their session; this page collects everything in one place. (Draft — citations will be finalized, and several arXiv items may have newer versions by Fall 2026.)

Books

• Ross Anderson, Security Engineering, 3rd ed. (Wiley, 2020). Chapters 1 (What is Security Engineering?), 4 (Access Control), and 21 (Network Attack and Defense). Author’s free copy.
• Florian Tramèr et al., Adversarial Machine Learning (Cambridge, 2024).

Video

• 3Blue1Brown, Neural Networks series. Ch. 3, “Backpropagation, intuitively” and Ch. 4, “Backpropagation calculus”. Watch them in that order: Ch. 3 gives you the picture of what the algorithm is doing to the weights, Ch. 4 derives the chain rule that does it. Between them they cover the backpropagation slide from the Week 2 deck at a much slower pace, with the animation doing the work a static figure cannot.

Foundations & classical offense

• Bell & LaPadula, “Secure Computer Systems: Mathematical Foundations” (MITRE, 1973).
• Shoshitaishvili et al., “SoK: (State of) The Art of War: Offensive Techniques in Binary Analysis” (IEEE S&P, 2016).
• Aleph One, “Smashing the Stack for Fun and Profit” (Phrack 49, 1996).
• Shacham, “The Geometry of Innocent Flesh on the Bone: Return-into-libc without Function Calls” (ACM CCS, 2007).

Automated exploit generation

• Cadar, Dunbar & Engler, “KLEE: Unassisted and Automatic Generation of High-Coverage Tests…” (OSDI, 2008).
• Avgerinos et al., “AEG: Automatic Exploit Generation” (NDSS, 2011).
• Cha et al., “Unleashing Mayhem on Binary Code” (IEEE S&P, 2012).

AI agents for security

• Yang et al., “SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering” (arXiv:2405.15793, 2024).
• Fang et al., “LLM Agents Can Autonomously Exploit One-Day Vulnerabilities” (arXiv:2404.08144, 2024) — read alongside published critiques.
• Xu et al., “When LLMs Meet Cybersecurity: A Systematic Literature Review” (Cybersecurity, Springer, 2025).
• Sheng et al., “LLMs in Software Security: A Survey of Vulnerability Detection Techniques and Insights” (ACM Computing Surveys, 2025).

Adversarial ML & network defense

• Mirsky et al., “Kitsune: An Ensemble of Autoencoders for Online Network Intrusion Detection” (NDSS, 2018).
• Goodfellow, Shlens & Szegedy, “Explaining and Harnessing Adversarial Examples” (ICLR, 2015).
• Carlini & Wagner, “Towards Evaluating the Robustness of Neural Networks” (IEEE S&P, 2017).

Social engineering

• Schmitt & Flechais, “Digital Deception: Generative Artificial Intelligence in Social Engineering and Phishing” (AI Review, Springer, 2024).

Attacks on AI systems

• Greshake et al., “Not What You’ve Signed Up For: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection” (AISec @ CCS, 2023).
• Zou et al., “Universal and Transferable Adversarial Attacks on Aligned Language Models” (arXiv:2307.15043, 2023).
• OWASP, “Top 10 for LLM Applications” (2025 edition).
• Nasr, Carlini, Sitawarin, Schulhoff et al., “The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against LLM Jailbreaks and Prompt Injections” (arXiv:2510.09023, 2025).
• Panfilov, Schaeffer, Schmotz, Shumailov, Beurer-Kellner, Prabhu, Geiping & Andriushchenko, “Stealing Reasoning Traces from Proprietary LLM APIs” (2026) — stolen-thoughts.com. Encrypted client-side chain-of-thought blocks are portable across sessions, users, and sibling models; a “decryption jailbreak” turns a weaker model into an oracle that outputs a stronger model’s reasoning verbatim, enabling distillation, mass PII/credential extraction from public logs, hazard leakage past refusals, and invisible prompt injection. Case study in the LLM & Agent Escapes deck (Week 12).

Tools & environments

• Ghidra — NSA’s open-source reverse-engineering suite.
• pwntools — CTF/exploit-development framework.
• angr — binary analysis & symbolic execution.
• Course sandbox / CTF range — details distributed in Week 1. All offensive work happens here, with authorization.

Slides

Decks are linked from each session on the schedule, and go up as we reach each topic. In any deck, press s for speaker notes, f for fullscreen, and Esc for the slide overview.