n8n-nodes-kybershield
v0.1.0
Published
n8n community node for KyberShield — drop a security check in front of any AI step in your workflow.
Maintainers
Readme
n8n-nodes-kybershield
KyberShield AI security as a drop-in n8n node. Add it between any AI step (OpenAI, Anthropic, HTTP request to an LLM) and the rest of your workflow to:
- Detect prompt injection / jailbreak attempts before they reach the model
- Validate AI tool / function calls against a policy allowlist
- Redact PII & secrets from LLM responses before they're sent to chat / email / Slack
- Sanity-check SQL queries generated by an AI agent
Install
In your n8n instance: Settings → Community Nodes → Install → n8n-nodes-kybershield
Configure
- Get your API key from the KyberShield portal under Agents.
- In n8n: Credentials → New → KyberShield API → paste the key.
Use
Drop a KyberShield node into any workflow. Pick an operation:
| Operation | Use it for | |---|---| | Check Prompt | User input that's about to be sent to an LLM | | Check Tool Call | An AI agent's proposed function/tool invocation | | Check Response | Redact PII/secrets from an LLM's raw output | | Check SQL Query | A SQL string generated by an AI agent |
The node has two outputs: Allowed and Blocked. Wire them however you want — e.g. send blocked items to a Slack alert and allowed items onward to the next AI step.
License: Apache-2.0
