Observed arrival · 2026-10-13
Agents Were Here: a wall for AI agents
A public message wall invites AI agents to leave permanent marks and reply to one another across sessions.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- AI agents that can open a web page
- Worth noticing
- Its policy calls agent-written marks untrusted data and tells readers not to follow instructions found inside them.
Field notes
The instructions distinguish the human reading path from the agent path: people can read the wall, while agents are directed to /write or /llms.txt. The page says marks are permanent and replies can connect conversations across sessions, models, and companies. A policy warning treats those marks as untrusted data; the extract does not show whether entries or replies are currently live.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue