A personal reading list on AI security
AI Security Corpus
One person's reading list that got out of hand. No gatekeeping, added and not endorsed: 5,660 links across 29 categories and 108 sections, sorted by what they are about.
Results appear below as you type. Tab or the arrow keys move through them, Enter opens one, Escape closes the list.
- 5,660links
- 29categories
- 108sections
- 1,155GitHub repos
- 570arXiv papers
- 570posts on X
Recently added 12 links, 2026-09-04
- @office-agents/sdk - headless agent runtime SDK for AI-powered Office Add-insnpmjs.com
- Agent Development Kit - ADK - build multi-agent systemsadk.dev
- agent-browser - browser automation CLI for AI agentsgithub.com
- agents-cli - build AI agents on Google Cloudgithub.com
- ANUS - Autonomous Networked Utility System open-source AI agent frameworkgithub.com
- Azure Functions serverless agents runtimelearn.microsoft.com
- Centaur - self-hosted control plane for sandboxed, credential-safe AI agents in Slackcentaur.run
- CLI-Anything - making all software agent-nativegithub.com
- CrewAI - multi-agent platformcrewai.com
- Cua launches stable Agent - computer-use agent framework for macOS with sandbox and VM isolationx.com
- DeerFlow - open-source long-horizon SuperAgent harnessgithub.com
- Eko - build production-ready agentic workflows with natural languagegithub.com
The corpus in numbers
About
Every link lives in a markdown file in the repository, one file per category, one heading per section. This site is generated from those files on every push. Nothing here is a recommendation: a link is in the corpus because it was worth reading once, not because it is right.
To add something, edit the category file and open a pull request, or open an issue with the URL. Titles are kept as the source wrote them. Licensed CC BY 4.0.