agent-cordon is an open-source Python package that provides a quarantine line for data ingested by LLM agents. It scans untrusted text and outbound actions for prompt injection and exfiltration risks, helping developers secure AI agent workflows. Designed for AI developers and security-conscious teams.
agent-cordon sits in PulseGate's Other AI category. It focuses on preventing prompt injection and data exfiltration in LLM agent workflows by scanning untrusted text and actions. It is built as an open-source project for AI developers and security engineers. agent-cordon is open source under the MIT license. agent-cordon is available on the command line.
agent-cordon first shipped in 2026. PulseGate's similarity index finds few close equivalents — agent-cordon occupies a relatively distinct niche. Key capabilities include prompt injection scanning, outbound action monitoring, and text data quarantine. It exposes integrations via an MCP server.
Latest indexed changes and source events
agent-cordon discovered by the PulseGate indexer
Other apps tracked under the same category.
