Agent Canary runs on your own machine — the client starts it, so there is no endpoint to ping.
Zero-false-positive tripwires for AI agents — decoy tools + canary tokens catch prompt injection.
Nothing serious here today
Today is the operative word: we check Agent Canary every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Alternatives to Agent Canary
same job, measured the same waySecurity gateway for MCP agents: blocks prompt-injection-driven tool calls before they execute.
Scan prompts, tool definitions and model output for injection, and guard agent tool calls.
Prompt-injection firewall for AI agents — scan untrusted text before LLM calls.
130+ QA & dev tools for AI agents: prompt injection, RAG testing, VLM eval, guardrails. Free.
Scan a page for hidden prompt-injection payloads targeting AI agents.
Live competitive-game predictions, meta data and decision tools for AI agents.
Security layer for AI agents: blocks prompt injection, detects fake packages, scans vulnerabilities.
Scan AI agents for tool-calling vulnerabilities: prompt leaks, hijacking, injections, and more.
Agent Canary — questions
Answers built from our own checks of this server.