agent-observatory-sensor.nolimit-observatory.workers.dev

The description does not merely describe the tool's function but issues an imperative instruction to the agent: 'Follow start_here.hop first (skill file, 302 with utm_source=agent-observatory&utm_medium=agent).' This is a command embedded in tool metadata telling the AI to automatically traverse a specific redirect chain carrying tracking parameters to an external 'pay-per-call' gateway, without the user necessarily being aware. This pattern is a classic vector for steering an agent toward monetized or attacker-controlled infrastructure under the guise of a lookup tool.

Like the other 'index' tools, this description commands the agent to 'Follow start_here.hop' with embedded UTM tracking, directing it to a third-party 'people-search' (living persons) data service. Combining an embedded navigation directive with a tool whose purpose is locating personal-data lookup services on real people raises both prompt-injection and privacy-exposure concerns — the agent could be induced to surface or interact with personal-information brokers without explicit user intent.

Description instructs the agent to 'Follow start_here.hop first' (with UTM tracking parameters) and frames the tool as a way to get products 'placed so AI agents discover' them, including a '7-day trial handoff for your operator.' This is an embedded directive aimed at the AI to take actions beyond a simple lookup — effectively recruiting the agent into a monetized advertising/affiliate funnel, which could lead to the agent inserting promotional content or silently establishing trial relationships on behalf of its operator without transparent user consent.