A Generative Security Application Engineering Curriculum

David Baker-Robinson; Wu-chang Feng

REVIEW 3 major objections 5 minor 57 references

A Generative Security Application Engineering Curriculum

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proposes a curriculum that teaches cybersecurity by building, attacking, and applying LLM applications.

desk verdict Solid, clearly written curriculum blueprint for LLM security, but the teaching is unvalidated and the public materials are not yet reproducible by outsiders. read the letter →

arxiv 2501.10900 v1 pith:EHSSQXZC submitted 2025-01-18 cs.CY cs.AI

Wu-chang Feng , David Baker-Robinson This is my paper · ORCID

classification cs.CYcs.AI

keywords generativeAIsecurityeducationLLMapplicationscurriculumdesignpromptinjectionretrieval-augmentedgenerationagentscapture-the-flag

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes a curriculum and a course that teaches cybersecurity by having students build, attack, and apply large language model (LLM) applications. Its central argument is that security education should stop treating generative AI as a topic and start treating it as the working medium: students learn the models, assemble them into applications, compromise those applications, and then use them on security tasks with known correct answers. The course is organized into four modules—models, frameworks, security, and tasks—and all materials are released publicly so other programs can reuse them. The intended payoff is a generation of security practitioners who can steer automation instead of performing every step by hand.

What carries the argument

The load-bearing mechanism is the four-part course architecture: models, frameworks, security, and tasks. Each segment supplies the next: students probe models to understand what they can and cannot do; assemble models into applications using a framework's abstractions for retrieval, agents, and tools; break those applications to learn LLM-specific failure modes; and finally use the resulting skills on security tasks whose outcomes are known and checkable. The repeated design motif is a 'naive versus hardened' pair—for example, a database agent with excessive tool privilege is contrasted with one that restricts permitted SQL and validates inputs—so that a security principle is learned by building and breaking. The ground-truth requirement is what keeps the course honest about LLM reliability: because models can be confidently wrong, every task in the final module has a correct answer students can verify.

What would settle it

A controlled teaching experiment: give one cohort this four-module curriculum and another a traditional security course, then give both new security tasks with known correct answers and measure accuracy, speed, and judgment about when to trust model output; if the LLM-curriculum group does not outperform, the central premise fails. A simpler check is to collect student solutions from the labs and see whether graduates of the course can independently detect prompt injection and restrict agent tool privileges when asked to build a fresh agent.

Watch

Extended reading notes

Core claim

On its own terms, the paper's claim is that a security course can be built around the generative-AI application life cycle and that doing so teaches students both how to use LLMs and how to secure them. The curriculum starts with hands-on model experimentation, then moves to building LLM applications with retrieval-augmented generation and agent architectures, then to attacking those applications using a standard catalog of LLM-specific vulnerabilities such as prompt injection and excessive agency, and finally to applying the tools to security tasks—code summarization, vulnerability discovery and exploitation, command and configuration generation, code generation, threat intelligence, and social engineering. A key design choice is the use of 'ground-truth' tasks, drawn from capture-the-flag levels and known commands and configurations, so students can verify whether a model's output is actually correct. The paper's contribution is the curriculum structure and the public materials, not an empirical demonstration of improved learning.

Load-bearing premise

The curriculum's value rests on the assumption that the hands-on exercises translate into real ability to use generative AI in security work, yet the paper reports no assessments, pre/post tests, or comparisons to establish that.

Editorial extensions

If this is right

Adopting this structure lets a security program teach LLM application development and security together rather than as separate electives.
Students completing the course should be able to build LLM-powered tools for code analysis, vulnerability discovery, command generation, threat intelligence, and social engineering detection.
The ground-truth task design gives students a way to benchmark whether a model is dependable, countering the danger of plausible-but-wrong outputs.
By making the materials public, the paper enables other institutions to offer a similar course without designing exercises from scratch.
The 'secure by construction' pattern—using custom restricted tools instead of broad agent toolkits—gives students a concrete template for secure LLM application design.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper's implicit pedagogical bet is that learning to build and break an LLM application transfers to real security work; the obvious next test is a controlled comparison of this course against a conventional security curriculum on new, unseen tasks.
The ground-truth benchmarking idea could be lifted out of the classroom and used as an evaluation methodology for LLM security tools generally, giving teams a cheap way to decide which models they can trust.
A testable extension is to measure not just whether students answer tasks correctly but whether they learn when to distrust model output, since the curriculum's emphasis on verifiable tasks should sharpen calibration.
The course implicitly argues that future security practitioners will spend more time designing and validating automated workflows than performing individual commands; if true, curriculum hours should shift away from memorizing tools and toward evaluation and tool-building.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick read of Feng and Baker-Robinson's curriculum paper. The honest headline: this is a solid, well-organized teaching blueprint, and it earns a place in the conversation about how to teach LLM security. It is not a research paper and should not be judged as one.

What's new: the specific course design. The sequence—model familiarity, RAG and agent labs, LLM security attacks, then ground-truth security tasks—is a useful assembly of known pieces (LangChain, DSPy, OWASP Top 10 for LLMs, Portswigger levels, CTFs) into a coherent semester. I especially like the use of tasks with known correct answers (Section 3.4), which forces students to wrestle with LLM hallucination instead of just chaining demo outputs. The Bulbasaur ReAct injection write-up is a nice teaching example. The authors also acknowledge model limitations and costs, which is more honest than most LLM course descriptions.

The soft spots, in order. First, no learning assessment. The course ran twice (Spring/Fall 2024) but no student outcomes, pre/post tests, or even enrollment attrition. The paper itself calls it an "initial curriculum," which is honest, but the load-bearing claim—that these exercises teach students to apply generative AI in security—is plausible and untested. Second, the reusability claim is oversold. Section 4 says the materials are "available for public use and modification," which is literally true (links to slides, labs, code), but the stress-test note about missing dependencies is fair: external adopters need API keys, billing info, and tool-specific setup that the paper does not document. That gap makes independent evaluation of the materials harder, not impossible. I'd downgrade the "help similar efforts" goal to "helpful but needs setup documentation." Neither flaw is fatal to the paper's descriptive purpose.

The citation pattern looks fine: references are mostly online docs, CTF portals, and prior LLM security work, appropriate for a curriculum piece. No citation inflation or self-citation abuse.

Bottom line: this is a useful design document for educators building or revising an LLM security course. It deserves serious peer review, but the expected outcome should be a revision that adds an evaluation plan and a short "getting started" appendix covering setup and costs. I'd bring it to a reading group as a discussion piece on what counts as evidence in CS education. Not my area to cite directly, but I'd point colleagues to it.

Referee Report

3 major / 5 minor

Summary. This paper describes a graduate-level or upper-division security course at Portland State University that teaches students how to apply generative AI and large language models to cybersecurity tasks. The curriculum is organized into four modules: LLM models and their characteristics; application frameworks, primarily LangChain, RAG, and agent architectures; security of LLM applications, including OWASP Top 10 vulnerabilities, prompt injection, and excessive agency; and a set of security tasks with verifiable outcomes, including code summarization, vulnerability discovery and exploitation, command and configuration generation, code generation, threat intelligence, and social engineering. Section 4 reports that versions of the course ran in Spring 2024 and Fall 2024 and that the course materials are publicly available, with links to slides, labs, source code, screencasts, and tutorial videos.

Significance. If the materials work as described, the paper is a useful, unusually concrete blueprint for an emerging area of security education. Its modular organization, detailed exercise descriptions, and inclusion of tasks with known correct answers are strengths; the RAG-based prompt injection exercise in §3.3 and the agent-securing exercises in §3.2.2 are specific enough for other instructors to understand and adapt. The authors appropriately call the work an initial curriculum rather than a validated one. However, the contribution as presented is a course description rather than an educational intervention: the paper reports no learning outcomes, no assessment data, and no external adoption evidence, and the public materials are not accompanied by reuse documentation. These omissions directly affect the paper's stated goal of helping similar efforts.

major comments (3)

[4. Conclusion] The central pedagogical claim is unsupported by assessment data. Section 4 states that the course was offered in Spring 2024 and Fall 2024, but the paper reports no learning outcomes, student work samples, pre/post measures, or comparison group, and no section evaluates whether the exercises achieve their stated learning goals. Without such evidence, the reader cannot distinguish an effective curriculum from a collection of demonstrations. Please add an evaluation, however preliminary, or explicitly reframe the paper as a descriptive curriculum proposal with evaluation left to future work.
[4. Conclusion and §3.1–§3.4] The public reuse claim is undermined by missing dependency and credential documentation. The conclusion says the materials are available 'for public use and modification' to 'help similar efforts,' but the exercises depend on Ollama/cloud VMs (§3.1), LangChain tools such as SerpAPI and arbitrary OpenAPI endpoints (§3.2.2), Portswigger and CTF services (§3.3 and §3.4.2), and multiple account-gated or paid APIs such as VirusTotal, crt.sh, Safe Browsing, PhishTank, and OOPSpam (§3.4.5). The linked repositories are not accompanied in the paper by a dependency manifest, API-key provisioning guide, cost estimate, or hardware requirement, so an external instructor cannot determine how to run the course. Please provide a run-book or installation guide, or soften the reuse claim to match what is actually provided.
[3.4.6 Social engineering] The social engineering module lacks an explicit safety and ethics framework. Students are asked to generate fake social media profiles, phishing lures, and misinformation, but the paper does not describe safeguards, platform terms-of-service considerations, human-subjects review, or how the generated artifacts are contained and disposed of. For a curriculum intended for adoption elsewhere, this omission is consequential; the authors should state the controls used in their offering, such as sandboxed environments, fictional targets, and institutional review procedures.

minor comments (5)

[3.3 and References [38]] The text and reference list use 'OW ASP' with a space; this should be 'OWASP'.
[3.4.6] The word 'evaulate' in the phishing lures exercise should be 'evaluate'.
[3.4.4 Chatbot coding] The acronym 'MVP' is defined as 'model-view-controller'; this is normally abbreviated 'MVC', and the text should use the correct abbreviation.
[3.4.3 Cloud commands] 'A WS' in 'A WS's aws' should be 'AWS'.
[3.4.1 Summarization] The rendering 'Pok´ emon' in the RAG prompt injection walkthrough has broken encoding; it should read 'Pokémon'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the paper is a descriptive curriculum report with no derivation chain, fitted parameters, or predictions to reduce to inputs.

full rationale

The paper describes a generative-AI security course and its materials; it contains no formal derivation, no equations, no fitted parameters, and no quantitative prediction. Its central sentence, 'we describe an initial curriculum and course that attempts to show students how to apply generative AI in order to solve problems in security,' is a statement of scope rather than a derived claim. The course content is assembled from external references such as LangChain, OWASP, Portswigger, and CTF exercises, and the paper's contribution is the organization and presentation of those materials. The self-references in the references section point to the authors' own slides, labs, code, and screencasts, but these are the artifacts being announced, not load-bearing evidence for an argument; no conclusion depends on a prior result by the same authors. No 'uniqueness theorem' or ansatz is imported from the authors' prior work, and no task is renamed as a prediction. The absence of learning assessments or external reproducibility evidence is a real limitation, but it is an evidence gap about effectiveness and adoptability, not a circularity. Therefore the paper's derivation chain, such as it is, is self-contained and no circular step exists.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim is a course description; it does not rest on fitted parameters or invented entities. It does assume domain premises: LLMs are changing security practice, the tools listed are appropriate for teaching, and hands-on exercises build relevant skills. No postulates beyond those assumptions appear.

assumptions (3)

domain assumption Generative AI and LLMs are materially changing cybersecurity practice.
Introduction, paragraph 1. The whole motivation depends on this trend, asserted by citation to external sources rather than demonstrated in the paper.
domain assumption Hands-on exercises with LLM frameworks transfer to real-world security skills.
Section 3 and Conclusion. The course's value claim assumes active labs improve students' ability to apply LLMs, but no student outcome data are reported.
domain assumption The specific tools (LangChain, Ollama, ChromaDB, DSPy) are appropriate and stable enough for teaching.
Sections 3.2 and 3.4. Tool choices rest on popularity and availability as of 2024; fast-moving versions could make materials date quickly.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Generative Security Application Engineering Curriculum." pith.science (2026). https://pith.science/paper/EHSSQXZC

@misc{pith2026250110900,
  author       = {Pith},
  title        = {Pith review of: A Generative Security Application Engineering Curriculum},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EHSSQXZC}},
  note         = {Machine review of arXiv:2501.10900}
}

read the original abstract

Generative AI and large language models (LLMs) are transforming security by automating many tasks being performed manually. With such automation changing the practice of security as we know it, it is imperative that we prepare future students for the technology landscape they will ultimately face. Towards this end, we describe an initial curriculum and course that attempts to show students how to apply generative AI in order to solve problems in security. By refocusing security education and training on aspects uniquely suited for humans and showing students how to leverage automation for the rest, we believe we can better align security education practices with generative AI as it evolves.

Figures

Figures reproduced from arXiv: 2501.10900 by the authors.

**Figure 1.** Course overview ent ways to interact with them, and many different tasks they can be applied towards. The curriculum begins with a short crash course on how models work including the probabilistic and non-deterministic underpinnings of LLMs, along with the ideas behind instruction-tuned models and the mixture of experts approach that enables more purpose-built responses from them. Common parameters that can be used… view at source ↗

**Figure 2.** Vulnerable SQL agent the context in order to improve accuracy and reduce model usage costs [44]. • Embedding models that can support similarity searches by representing semantic content in documents as a large, multi-dimensional vector [18]. • Vector databases and indexes that can store documents and support similarity searches across them such as ChromaDB [5]. • Chains that can concisely compose the above compon… view at source ↗

**Figure 3.** Securing LLM agents with custom tools experiment with LangGraph, a framework built on top of LangChain, which enables the creation of LLM applications using a graph-based control flow [27]. Students use LangGraph to experiment with the customization of AI agents and the trade-offs between control and agency. At the end of the course, students experiment with DSPy, a framework that attempts to shift developers away … view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: ReAct prompt template injection locations [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Prompt injection of ReAct template • Deobfuscation: For reverse engineers, the code and data being analyzed may be obfuscated via hex-encoding, base64 encoding and XORbased encryption to evade signature detection and make reverse engineering more difficult. In this pa…

**Figure 6.** Figure 6: CTF level DOM-based cross-site scripting flaws in them. • Memory corruption vulnerabilities: Memory corruption vulnerabilities are common in unmanaged languages such as C. In this part, students are given common memory corruption flaws such as buffer overflow, out-of…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 56 canonical work pages

[1]

Ackerman, M

D. Ackerman, M. Chakrabarti, and T. Skoog. Pig Butchering: The online scam that’s raked in $75 billion and counting, 2024. https://www. wbur.org/onpoint/2024/04/03

work page 2024
[2]

Dear diary: A randomized controlled trial of generative ai coding tools in the workplace

Jenna Butler, Jina Suh, Sankeerti Haniyur, and Constance Hadley. Dear diary: A randomized controlled trial of generative ai coding tools in the workplace. arXiv preprint arXiv:2410.18334, 2024

arXiv 2024
[3]

Free Email Address Validation API Service, 2024

Check-Mail. Free Email Address Validation API Service, 2024. https://check-mail.org/

work page 2024
[4]

Combating Misinfor- mation in the Age of LLMs: Opportunities and Challenges, 2023

Canyu Chen and Kai Shu. Combating Misinfor- mation in the Age of LLMs: Opportunities and Challenges, 2023

work page 2023
[5]

Chroma, 2024

Chroma. Chroma, 2024. https://www. trychroma.com/

work page 2024
[6]

It’s True, LLMs are Better Than People at Creating Convincing Misinfor- mation, 2024

Thomas Claburn. It’s True, LLMs are Better Than People at Creating Convincing Misinfor- mation, 2024. https://www.theregister.com/ 2024/01/30/llms_misinformation_human

work page 2024
[7]

EV A Email Verification API, 2024

Compile7. EV A Email Verification API, 2024. https://eva.pingutil.com/

work page 2024
[8]

Prompt Engineering Guide, 2024

DAIR.AI. Prompt Engineering Guide, 2024. https://www.promptingguide.ai/

work page 2024

Show all 57 references

[9]

Welcome to Pydantic,

Pydantic Developers. Welcome to Pydantic,
[10]

W. Feng. 2024 NSF SaTC PI Meeting Tutorial notebooks. https://bit.ly/pdx-satc24

2024
[11]

W. Feng. Portland State University Generative Security Applications code. https://github. com/wu4f/cs410g-src
[12]

W. Feng. Portland State University Gen- erative Security Applications Spring 2024 Screencasts on YouTube. https://bit.ly/ pdx-gensec-videos

2024
[13]

aider, 2024

Paul Gauthier. aider, 2024. https://aider. chat

2024
[14]

GitHub Copilot: Your AI pair programmer, 2022

GitHub, Inc. GitHub Copilot: Your AI pair programmer, 2022. https://github.com/ features/copilot

2022
[15]

From Assistant to Analyst: The Power of Gemini 1.5 Pro for Malware Anal- ysis, April 2023

Google. From Assistant to Analyst: The Power of Gemini 1.5 Pro for Malware Anal- ysis, April 2023. https://cloud.google. com/blog/topics/threat-intelligence/ gemini-for-malware-analysis

2023
[16]

Google Safe Browsing, 2024

Google. Google Safe Browsing, 2024. https: //developers.google.com/safe-browsing

2024
[17]

Supercharge Security with AI, 2024

Google. Supercharge Security with AI, 2024. https://cloud.google.com/security/ai

2024
[18]

Text Embeddings API, 2024

Google. Text Embeddings API, 2024. https://cloud.google.com/vertex-ai/ generative-ai/docs/model-reference/ text-embeddings-api

2024
[19]

PentestGPT: A GPT-empowered Penetration Testing Tool, 2023

GreyDGL. PentestGPT: A GPT-empowered Penetration Testing Tool, 2023. https:// github.com/GreyDGL/PentestGPT

2023
[20]

PhishTank: Join the Fight Against Phishing, 2024

Cisco Talos Intelligence Group. PhishTank: Join the Fight Against Phishing, 2024. https:// phishtank.org/

2024
[21]

A Survey on Retrieval-Augmented Text Generation for Large Language Models, 2024

Yizheng Huang and Jimmy Huang. A Survey on Retrieval-Augmented Text Generation for Large Language Models, 2024

2024
[22]

IP Geolocation API and IP Location Lookup Tools, 2024

ipwhois.io. IP Geolocation API and IP Location Lookup Tools, 2024. https://ipwhois.io/

2024
[23]

Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christopher Potts

Omar Khattab, Arnav Singhvi, Paridhi Mahesh- wari, Zhiyuan Zhang, Keshav Santhanam, Sri Vardhamanan, Saiful Haq, Ashutosh Sharma, Thomas T. Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christopher Potts. DSPy: Compiling Declarative Language Model Calls into Self-Imp...

2023
[24]

LangChain, 2024

LangChain. LangChain, 2024. https://www. langchain.com/

2024
[25]

LangSmith, 2024

LangChain. LangSmith, 2024. https://pypi. org/project/langchain-experimental/

2024
[26]

LangSmith, 2024

LangChain. LangSmith, 2024. https://www. langchain.com/langsmith

2024
[27]

Langgraph

LangChain, Inc. Langgraph. https://www. langchain.com/langgraph. Accessed: 2025-01- 18

2025
[28]

Who’s Verifying the Verifier: A Case-Study in Securing LLM Applications, 2024

Dan Leininger. Who’s Verifying the Verifier: A Case-Study in Securing LLM Applications, 2024. https://innovation.consumerreports.org

2024
[29]

LlamaIndex, Data Framework for LLM Applications, 2024

LlamaIndex. LlamaIndex, Data Framework for LLM Applications, 2024. https://www. llamaindex.ai/

2024
[30]

Microsoft Threat Intelligence Se- curity Tools, 2024

Microsoft. Microsoft Threat Intelligence Se- curity Tools, 2024. https://github.com/ microsoft/msticpy

2024
[31]

Introducing Mi- crosoft Security Copilot, 2023

Microsoft, Inc. Introducing Mi- crosoft Security Copilot, 2023. https: //www.microsoft.com/en-us/security/ business/ai-machine-learning/ microsoft-security-copilot

2023
[32]

Large Language Mod- els and Cybersecurity: Proceedings of a Workshop - in Brief, 2023

National Academies of Sciences, Engineer- ing, and Medicine. Large Language Mod- els and Cybersecurity: Proceedings of a Workshop - in Brief, 2023. https://nap. nationalacademies.org/read/27776

2023
[33]

Ghidra Software Reverse Engineering Suite, 2024

National Security Agency. Ghidra Software Reverse Engineering Suite, 2024. https:// ghidra-sre.org/

2024
[34]

Ollama, 2024

Ollama. Ollama, 2024. https://ollama.com/

2024
[35]

OOPSpam Anti-Spam API, 2024

OOPSpam. OOPSpam Anti-Spam API, 2024. https://www.oopspam.com/

2024
[36]

OpenCVE, 2024

opencve.io. OpenCVE, 2024. https://www. opencve.io/

2024
[37]

OverTheWire Wargames, 2024

Overthewire. OverTheWire Wargames, 2024. http://overthewire.org/

2024
[38]

OW ASP Top 10: LLM Applications and Generative AI, 2024

OW ASP Foundation. OW ASP Top 10: LLM Applications and Generative AI, 2024. https: //genai.owasp.org/

2024
[39]

Risky Business Episode 709, June 2023

Patrick Gray and Adam Boileau and Marco Slaviero. Risky Business Episode 709, June 2023. https://risky.biz/RB709/

2023
[40]

Web LLM Attacks, 2024

Portswigger. Web LLM Attacks, 2024. https://portswigger.net/web-security/ llm-attacks

2024
[41]

Web Security Academy, 2024

PortSwigger. Web Security Academy, 2024. https://portswigger.net/web-security

2024
[42]

ATTACK Python Client,

Roberto Rodriguez. ATTACK Python Client,
[43]

Generative AI Security Adventures, 2024

Roberto Rodriguez. Generative AI Security Adventures, 2024. https://otrf.github.io/ GenAI-Security-Adventures

2024
[44]

https://attackcti.com/
[45]

A Systematic Survey of Prompt Engi- neering in Large Language Models: Techniques and Applications, 2024

Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, and Aman Chadha. A Systematic Survey of Prompt Engi- neering in Large Language Models: Techniques and Applications, 2024

2024
[46]

Chunk Visualizer, 2024

Aymeric Roucher. Chunk Visualizer, 2024. https://huggingface.co/spaces/m-ric/ chunk_visualizer

2024
[47]

SerpApi, 2024

SerpApi. SerpApi, 2024. https://serpapi. com/

2024
[48]

crt.sh Certificate Search, 2024

Sectigo. crt.sh Certificate Search, 2024. https: //crt.sh/

2024
[49]

Poisoned LangChain: Jailbreak LLMs by LangChain, 2024

Ziqiu Wang, Jun Liu, Shengkai Zhang, and Yang Yang. Poisoned LangChain: Jailbreak LLMs by LangChain, 2024

2024
[50]

VirusTotal - Home, 2024

VirusTotal. VirusTotal - Home, 2024. https: //virustotal.com/

2024
[51]

Generative Security Ap- plications Slides, 2024

Wu-chang Feng. Generative Security Ap- plications Slides, 2024. https://bit.ly/ pdx-gensec-slides

2024
[52]

Generative Security Applica- tion Engineering

Wu-chang Feng. Generative Security Applica- tion Engineering. In 2024 NSF Workshop on LLMs for Network Security , 2024. https:// www.youtube.com/watch?v=CTRGYAA-MtE

2024
[53]

Generative Security Applications Labs, 2024

Wu-chang Feng and David Baker-Robinson. Generative Security Applications Labs, 2024. https://codelabs.cs.pdx.edu

2024
[54]

Teaching Generative Security Applications using LangChain

Wu-chang Feng. Teaching Generative Security Applications using LangChain. In 2024 NSF SaTC PI Meeting Tutorial, 2024. https://www. youtube.com/watch?v=IdEIM7YbJDw

2024
[55]

Project Naptime: Evaluating Offensive Security Capabil- ities of Large Language Models, 2024

Google Project Zero. Project Naptime: Evaluating Offensive Security Capabil- ities of Large Language Models, 2024. https://googleprojectzero.blogspot. com/2024/06/project-naptime.html

2024
[56]

ReAct: Synergizing Reasoning and Acting in Language Models, 2023

Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. ReAct: Synergizing Reasoning and Acting in Language Models, 2023

2023
[2024]

https://docs.pydantic.dev/

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [1]

Ackerman, M

D. Ackerman, M. Chakrabarti, and T. Skoog. Pig Butchering: The online scam that’s raked in $75 billion and counting, 2024. https://www. wbur.org/onpoint/2024/04/03

work page 2024

[2] [2]

Dear diary: A randomized controlled trial of generative ai coding tools in the workplace

Jenna Butler, Jina Suh, Sankeerti Haniyur, and Constance Hadley. Dear diary: A randomized controlled trial of generative ai coding tools in the workplace. arXiv preprint arXiv:2410.18334, 2024

arXiv 2024

[3] [3]

Free Email Address Validation API Service, 2024

Check-Mail. Free Email Address Validation API Service, 2024. https://check-mail.org/

work page 2024

[4] [4]

Combating Misinfor- mation in the Age of LLMs: Opportunities and Challenges, 2023

Canyu Chen and Kai Shu. Combating Misinfor- mation in the Age of LLMs: Opportunities and Challenges, 2023

work page 2023

[5] [5]

Chroma, 2024

Chroma. Chroma, 2024. https://www. trychroma.com/

work page 2024

[6] [6]

It’s True, LLMs are Better Than People at Creating Convincing Misinfor- mation, 2024

Thomas Claburn. It’s True, LLMs are Better Than People at Creating Convincing Misinfor- mation, 2024. https://www.theregister.com/ 2024/01/30/llms_misinformation_human

work page 2024

[7] [7]

EV A Email Verification API, 2024

Compile7. EV A Email Verification API, 2024. https://eva.pingutil.com/

work page 2024

[8] [8]

Prompt Engineering Guide, 2024

DAIR.AI. Prompt Engineering Guide, 2024. https://www.promptingguide.ai/

work page 2024