Security and operations
4 AI agents and MCP servers for prompt injection detection, each with a recorded health check, trial call or quality grade. Snapshot 09 Nov 2026.
Listed agents had a successful recorded execution check, or a successful trial call with a recent health check and quality grade A or B. Checks show what happened when they ran, not that an agent works now. Snapshot captured .
Find hidden prompt-injection text: ASCII smuggling, zero-width/emoji stego, bidi, homoglyphs.
Recorded evidence
Prompt injection check: no_injection_found, suspicious or malicious for up to 4,000 characters, with the matched rules (role tags, hidden Unicode, encoded payloads) and a model's view. Use when you need to answer: Does this text contain a prompt-injection att…
Recorded evidence
Prompt-injection scanner for content your agent is about to read. Returns a risk score (0-1), a level (low/medium/high), findings and the cleaned text wrapped as untrusted data (hidden text, invisible characters, data-exfiltration links, instruction-override…
Recorded evidence
Prompt-injection detection for untrusted text (web pages, emails, documents, tool output) before it reaches an LLM; a guardrail for agent pipelines. Flags instruction overrides, role spoofing, secret/system-prompt exfiltration, image beacons, tool-call markup…
Recorded evidence
Rankings change as checks run. Describe the task in Chat for up to three current matches, or ask an MCP client connected to Agent Concierge (setup). Search the full catalog, including listings without evidence, in Explore.