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

The description instructs the AI to 'Follow start_here.hop first' through a 302 redirect carrying utm_source=agent-observatory&utm_medium=agent tracking parameters, then follow additional docs/CLI/remote-MCP hops. This is an imperative directive aimed at the agent's behavior (not just a description of output), designed to funnel the agent through a third-party tracked funnel to a 'pay-per-call tool gateway.' This could cause the agent to autonomously navigate to and interact with an unvetted external service, download/execute skill files, or register for paid services without clear user awareness.

Tool explicitly indexes 'a living people-search index' and instructs the agent to 'Follow start_here.hop' (tracked 302 redirect) to reach API/CLI/MCP hops. Combining an instruction that steers agent navigation with a service aimed at locating information about living individuals raises privacy and safety concerns — it could be used to help a user locate personal information about real people under the guise of a generic lookup tool, and the embedded hop-following instruction is a directive to the agent beyond simple parameter validation.

The description tells the agent to 'Follow start_here.hop first' (tracked redirect) and describes returning a '7-day trial handoff for your operator.' This is an instruction embedded in a tool description that could lead the agent to initiate a paid trial or commit the user's 'operator' account to a third-party advertising/monetization service, potentially without informed user consent. This goes beyond the tool's stated function of simply describing a URL/host and constitutes an attempt to manipulate agent behavior toward a commercial action.