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)
-
Taming SAST False Positives, the Adoption Killer
There is a predictable arc to a badly-run SAST deployment. The tool is switched on at full sensitivity, it returns thousands of findings.
-
TLS and PKI Hygiene: The Certificates You Forgot
TLS failures are rarely cryptographic. They are operational: expired certificates, copied private keys, undocumented internal CAs and shared wildcard certs.
-
Fuzzing Fundamentals: Letting the Machine Find the Weird Inputs
The inputs a developer tests are the inputs a developer imagines, and attackers make their living on the ones nobody imagined.
-
Service Accounts, Machine Identity and Secret Sprawl
Ask an organization how many employees it has and you get an exact number.
-
Shift-Left That Developers Actually Accept
Shift-left is the right idea wrapped in a phrase that has been ruined by overuse.
-
System Prompt Leakage: Secrets You Should Never Have Put There
Ask a chatbot the right way and it will often recite its own system prompt, the hidden instructions that define its persona, its rules, and its guardrails.
-
Tokenization and Data Masking: Making Data Useless to Steal
Encryption answers 'how do I protect data I need to use?' Tokenization and masking answer a sharper question.
-
CSPM: Finding Misconfigurations Before Attackers Do
CSPM scans your own cloud accounts continuously for the open bucket, the wildcard policy and the world-open security group, so you fix them first.
-
DDoS Resilience: Designing for the Flood
A distributed denial-of-service attack does not exploit a bug. It exploits arithmetic, more requests, or more raw bandwidth.
-
Evaluations and Red-Teaming as Security Controls
Once you accept that an AI system's behavior is a distribution rather than a fixed value, one question follows immediately: how do you ever get.
-
Image Signing and Provenance with Sigstore and cosign
Image scanning answers 'is this artifact vulnerable?' Signing and provenance answer a different, prior question: 'is this the artifact my pipeline.
-
Security Misconfiguration: Death by a Thousand Defaults
API security misconfiguration has no single signature: a permissive CORS policy, a debug endpoint left on, or a stack trace can each leave an API exposed.
-
Software Composition Analysis and the SBOM
Look at a modern application by line count and the uncomfortable truth is that you wrote very little of it.
-
Threat Hunting Fundamentals: Looking for What Your Alerts Missed
Every detection you own encodes an attack you anticipated. Threat hunting exists for the rest.
-
The EU AI Act: What Security Teams Actually Need to Do
Most coverage of the EU AI Act is written for lawyers and policy teams.
-
Vector and Embedding Weaknesses: The Security of RAG
Retrieval-augmented generation is how most enterprises make an LLM useful: embed your documents into vectors, store them, and retrieve the relevant.
-
A Logging Strategy That Pays Off in an Incident
There is a specific, sinking feeling in incident response: you have the attacker in your sights, you know roughly what they did, and the log that would.
-
Multi-Account and Landing-Zone Strategy for Blast-Radius Control
Ask where a cloud breach stops and the answer is usually 'at the edge of the account it happened in', if you are lucky enough to have edges.
-
Container Image Scanning and the SBOM
Software composition analysis for containers answers a deceptively simple question: what is actually inside this image, and which of those components.
-
Embeddings and the Vector-Database Threat Model
Retrieval-augmented generation gets the attention, but the layer underneath it, the embeddings and the vector database that stores them, has a threat.
-
Improper Inventory Management: You Can't Defend What You Can't See
Of all the OWASP API risks, this is the closest to our own focus: you cannot secure an AI endpoint you do not know exists.
-
MCP Security Architecture: Authentication, Scopes and Transport
The Model Context Protocol solved a real, boring problem: every agent needed its own bespoke glue to reach every tool and data source, and none of it.
-
Misinformation and Hallucination as a Security Risk
Hallucination is usually filed under 'model quality,' a problem for the data-science team. LLM09 insists it is also a security problem.
-
Purple Teaming: Turning Offense Into Detection
Purple teaming puts attackers and defenders together so each technique is either caught or becomes a detection you keep, not a line in a report.