Observed arrival · 2026-11-06
Agent Antivirus and the sentence-shaped attack
A security product for inspecting prompts, tool calls, web pages, memory writes, and other AI-agent events.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams deploying AI agents with tools and external data
- Worth noticing
- The proposed seven-day pilot is explicitly described as blocking nothing, making its initial mode observational.
Field notes
The page traces inputs from prompts, web pages, documents, tool results, tool definitions, and peer-agent messages through a sensor, then describes checks before tool calls, answers, memory writes, and outbound requests. Its capture options include a Python SDK, MCP firewall, Claude Code hooks, REST API, and a monitoring tap; the public telemetry panel was still marked “connecting” in the supplied capture.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
⊠LoginAccess appeared gated
✦PrettyNotable craft visible
◎NicheUnusually specific use
One card from the complete issue