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 & context management
Learn what the context window holds, how truncation, summaries and retrieval keep a prompt within it, and what each choice costs.
-
Tokens & tokenization
Understand how tokenizers turn text into integer IDs, and why tokens, not words, set your costs, limits and truncation.
-
Unstructured-data ETL for AI
Follow raw documents through parsing, cleaning, deduplication and chunking, and learn where each step breaks retrieval quality.
-
Vector Databases
See how a vector database indexes embeddings for approximate nearest-neighbor search, and when pgvector is enough instead.
-
Voice / Realtime API Agents
Compare a single speech-to-speech model with the STT, LLM and TTS chain, including latency targets, barge-in and what you lose in control.
-
Workflows vs Agents
Learn when to script an LLM process as a fixed workflow and when to let an agent choose its own steps, and what each costs.
-
Enterprise AI Security Assessment: A Practical Checklist
An AI security assessment answers five questions: what AI exists, who owns it, what it can reach, how it can be attacked, and what shows it is controlled.
-
How CISOs Can Build an Enterprise AI Security Inventory
Build the inventory from evidence, not from a survey. Combine what the endpoints, the cloud accounts, the identity system and the network can actually see.
-
What Is AI Security Posture Management (AI-SPM)?
AI security posture management (AI-SPM) is the practice, and the tooling, of continuously finding the AI an organization runs, working out how exposed.
-
The Confused Deputy Problem in AI Agents
The confused deputy is one of the oldest ideas in security: a program with legitimate authority is tricked by a less-privileged party into misusing.
-
What RAG Really Is, and Why You Need a Vector Database
A base language model knows only what it learned in training. It cannot see last night's support tickets, your internal wiki.
-
BOLA: The API Flaw That Tops the OWASP List
Change the number in the URL, get someone else's data. That one sentence describes Broken Object Level Authorization.
-
Broken Access Control: The Flaw That Now Tops the OWASP List
In the 2021 revision of the OWASP Top 10, Broken Access Control climbed to the number-one spot, overtaking injection after years at the top.
-
Securing the CI/CD Pipeline That Builds Your Images
A container image carries the trust of the pipeline that built it, and that pipeline holds credentials to push to registries.
-
Securing the CI/CD Pipeline: Every Stage Is an Attack Surface
The CI/CD pipeline is the most privileged system most organizations barely secure.
-
Sensitive Information Disclosure in LLM Applications
Of all the risks in the OWASP Top 10 for LLM applications, sensitive information disclosure is the one that rarely announces itself.
-
Shadow AI: The Invisible Attack Surface
Every security team has an asset inventory. Almost none of them can answer a simpler question: which AI tools, agents.
-
Broken Authentication: When the API Front Door Doesn't Lock
Authentication is the claim an API makes about who is on the other end of the connection.
-
Cryptographic Failures: When the Data Was Never Really Protected
The 2021 Top 10 renamed this category from 'Sensitive Data Exposure' to Cryptographic Failures, and the rename matters.
-
From Finding to Fix: Remediation Workflows That Stick
Every security program can produce findings. The ones that reduce risk can produce fixes, and the gap between the two is where most programs quietly fail.
-
Least Privilege in Practice, Not on a Slide
Least privilege is the security principle no one argues with and almost no one achieves.
-
Measuring an AppSec Program: Metrics That Matter
Ask most application security teams how their program is doing and you will hear a number of findings.
-
Network Segmentation That Actually Contains a Breach
Segmentation is the difference between an incident and a disaster. When an attacker lands on one machine, through a phished credential.
-
The Container Supply Chain: From Dockerfile to Registry to Runtime
A production container sits at the end of a chain of base images, Dockerfiles, registries and builds. Each handoff is a trust boundary attackers target.