About
AgentShield is a prompt injection classifier that detects jailbreaks, injection attacks, and adversarial inputs before they reach your model. Deploy as a cloud API or self-hosted Docker container. Context-aware classification, explainable verdicts, fail-open/closed policies, and Prometheus metrics built in. Free tier included.
Links
Badges

Maker History
🎉
Joined Product HuntApril 27th, 2026
Forums
View more

