Configure granular control flow and data flow rules

Attach guardrails to your agenticatic control and dataflows

Select LLM agent:

No Code After URL Fetch

Disallow code execution right after fetching a URL.

Code Vulnerability Scan

Scan generated Python/Bash for security issues.

GitHub-to-Pip Safety

Detect risky patterns in tool call sequences.

Excessive Code Smells

Check for ill-formed code using static code analysis

Secret leak detection

Make use of Semgrep for deep static code analysis of code

Email Restriction

Block emails to anyone except 'Peter' after viewing the inbox.

RAG Protection

Prevent unauthorized access to your RAG app.

Block PII

Scan for and block PII

Prompt Injection Guard

Spot and stop prompt injection in tool responses.

Link Trust Filter

Block untrusted links from tool outputs.

Harmful Content Filter

Stop processing of toxic or unsafe messages.