Ferro Labs Managed Plugins
Ferro Labs Managed adds five security plugins on top of the 6 open-source plugins that ship with the AI Gateway. They slot into the same before_request/after_request plugin pipeline, but the detection logic (PII models, secret patterns, injection scoring, schema/regex engines) is closed-source and only runs in managed deployments.
| Plugin | What it guards | Typical action |
|---|---|---|
pii-redact | Personally identifiable information in request or response content | Redact in place, or block the request |
secret-scan | Leaked API keys, credentials, and tokens in request content | Block, or warn and pass through |
prompt-shield | Prompt injection and jailbreak attempts in user messages | Block above a confidence threshold |
schema-guard | Model output conformance to a JSON Schema | Block non-conforming responses |
regex-guard | Custom regex-defined patterns (SSNs, business rules) | Block, or warn and pass through |
Ferro Labs Managed
These five plugins are not included in the open-source build — they are only available on Ferro Labs Managed. Join the waitlist for access.
Related
- Plugins overview — the 6 OSS plugins and the shared pipeline stages
- Ferro Labs Managed overview
- OSS vs. Ferro Labs Managed