Supply Chain Security, in plain English.

Sandboxing LLM Agent Code Execution: Patterns

If your agent can execute code, something it reads from the internet can execute code. Pick your sandbox before the agent picks one for you.

AI Model Weight Tampering Detection Techniques

Weight-level tampering leaves cryptographic and statistical fingerprints. Here is what current research says about detecting a modified checkpoint before it reaches inference.

MCP Server Capability Declaration Audit

An MCP server tells the world what it can do through its capability declaration. Auditing those declarations catches drift, tool poisoning, and misconfiguration before an agent gets talked into using the wrong one.

SSRF Detection: Griffin AI vs Mythos

Server-side request forgery is a test of how well your scanner understands the boundary between trusted and untrusted URLs. Griffin's engine resolves URL construction through string builders, template engines, and HTTP client configuration; Mythos reads the code and guesses. On modern applications that is the difference between a finding you can ship and a finding you cannot defend.

Citation Accuracy: Griffin AI vs Mythos

An AI security tool that cites the wrong advisory is worse than one that says nothing. Griffin AI benchmarks citation accuracy at 0.89 similarity; Mythos does not.

ChatGPT Plugins and the New Plugin Supply Chain Attack Surface

AI plugins connect LLMs to external services, creating a supply chain of trust that most users never examine. The risks are significant.

Cache Hit Optimisation: Griffin AI vs Mythos

Prompt caching and engine memoisation combine to make Griffin AI scans repeat-cheap. Pure-LLM tools recompute the same reasoning on every run.

Griffin AI vs AI21 Jurassic for Security Workflows

Griffin AI vs Qwen for Code Security

Qwen's open-weight models have strong code benchmarks. We dig into how they compare to Griffin AI when the workflow is real code security, not just leetcode.