Knowledge base
The AI and cloud security library
Practical, vendor-neutral explainers on AI, API, web, container and cloud security. All 207 are free to read, with no login. New here? Start with path 01, then work across.
Filter by topic (137)
-
The Context Window as a Trust Boundary
An LLM keeps trusted instructions and hostile text in the same context window, with no notion of privilege. That is the root of most prompt-injection stories.
-
Zero Trust Principles, Beyond the Buzzword
Few terms have been sold harder or understood less than Zero Trust. Strip away the marketing and it is a single.
-
Broken Object Property Level Authorization: Mass Assignment and Its Twin
The 2023 OWASP API list folded two well-known vulnerabilities into one category, and the merger makes sense once you see the shared root.
-
DAST: Dynamic Application Security Testing, Explained
There are two fundamentally different ways to look for flaws in an application: read the source, or attack the running thing.
-
Data Classification: The Foundation Everything Else Stands On
Ask a room of engineers to encrypt the sensitive data, and the first question back is which data that is.
-
IAM Misconfiguration: A Common Root Cause of Cloud Breaches
When a cloud breach gets written up, the headline usually names a leaked bucket or an exposed server.
-
Injection: SQL, XSS, and the Untrusted Input Problem
Injection was first on the OWASP Top 10 in 2010, 2013 and 2017. In 2021 it fell to third, behind Broken Access Control and Cryptographic Failures.
-
RBAC vs ABAC: Choosing an Access Model You Can Live With
Every access-control conversation eventually reduces to two verbs: be and have .
-
Runtime Security and Threat Detection for Containers
Every control up to this point is preventive: minimal images, signed provenance, admission gates, tight RBAC and network policy.
-
SAST for the AI Era: Scanning Prompts, Agents, and LLM Glue Code
Taint analysis transfers directly to AI applications, but prompts, model calls and agent tools add new sources and sinks that older SAST rulesets do not know.
-
Secrets in CI and How They Leak
The pipeline needs secrets to do its job, registry credentials, cloud keys, deploy tokens, signing keys.
-
Supply Chain Risk in the LLM Stack
LLM03 treats each borrowed piece of an AI system, from the base model to adapters and vector stores, as a link in a supply chain with AI-specific failure modes.
-
The AI Supply Chain: Models, Datasets, Adapters and Plugins
When you deploy an AI feature, it feels like one artifact: a model that answers questions.
-
The Lethal Trifecta: Private Data, Untrusted Content, and an Exfiltration Path
Simon Willison's lethal trifecta: an agent with private data, untrusted content and a way to communicate externally can be made to leak secrets.
-
Egress Control: Why Outbound Traffic Is the Blind Spot
Most firewalls describe what they block coming in and allow almost everything going out, yet every breach stage after initial access is outbound traffic.
-
Insecure Design: The Flaws You Cannot Patch
Insecure Design is a new category in the 2021 OWASP Top 10. It covers flaws in a system's design that no amount of careful coding can fix.
-
Lateral Movement in the Cloud After Initial Access
Lateral movement in a traditional network looks like an attacker hopping from machine to machine, leaving traffic a defender might spot.
-
Microsegmentation in Practice, Without Grinding to a Halt
The single most damaging property of most corporate networks is that they are flat: compromise one host and you can reach thousands.
-
Secrets Management in Kubernetes: etcd Encryption and External Stores
The single most common misconception in Kubernetes security is that a Secret object keeps a secret. It does not, not on its own.
-
SIEM and the Art of Useful Logging
Every security team eventually buys a SIEM, and a surprising number end up with an expensive log warehouse that detects nothing.
-
Triaging and Prioritizing Findings With Real Risk Context
Turn on enough scanners and you will drown. A large estate produces more findings than any team can fix.
-
Agent Identity and Delegated Authorization
An AI agent that does anything useful needs to act on systems: read a calendar, query a database, send a message.
-
Artifact Integrity, SLSA and Provenance
There is a gap in most software supply chains that hides in plain sight.
-
Cloud Logging and Detection: CloudTrail and Friends
The cloud gives defenders one large advantage: because every action is an API call, every action can be logged, in structured form, with the identity.