Free IDE extension for risk-free vibe coding — keep secrets out of AI.
Control
POST /api/intent/action/checkIntent verification system that ensures every agent action aligns with the original user's intent. Prevents goal hijacking, task drift, and multi-step manipulation attacks.
Complete these steps in order. The dashboard link opens the exact product workspace for this service.
Sign in, select the project you want to protect, then open Intent Guard.
When a user provides a task, Intent Guard extracts and encodes the core intent using semantic analysis.
Use POST /api/intent/action/check from server-side code. Keep API keys out of browsers and client bundles.
Run one safe case and one expected-risk case. Confirm the result in Intent Guard, then review Guard Logs before enabling the control for production traffic.
Advanced attacks don't try to inject at once — they gradually steer the agent away from the original task. Intent Guard tracks the original user intent and flags actions that diverge, catching sophisticated multi-turn manipulation attacks.
A step-by-step breakdown of how Intent Guard protects your AI.
When a user provides a task, Intent Guard extracts and encodes the core intent using semantic analysis.
Every action the agent takes is compared against the original intent vector.
Actions that significantly diverge from the original intent are flagged with a drift score.
Actions exceeding drift thresholds are blocked, questioned, or sent for human review.
Track user intent through multi-turn conversations and complex task chains
Quantify how much an action deviates from the original intent
Detect slow, subtle manipulation that happens over many turns
Set different drift tolerance levels based on action risk
This service is implemented at the application route below. Use the tested SDK and framework guides for copy-paste integration examples.
POST /api/intent/action/check. API keys must stay server-side. See the REST API guide and related documentation before integrating.