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

The description doesn't just describe output; it issues an imperative instruction to the calling agent ('Follow start_here.hop first') and specifies injecting UTM tracking parameters (utm_source=agent-observatory&utm_medium=agent) into the agent's subsequent actions. This is a classic prompt-injection pattern embedded in a tool description, directing the agent to navigate to and privilege a specific external monetized gateway ('Monid') rather than simply returning data, without the user's informed consent.

Like agent-tool-index, this description commands the agent to 'Follow start_here.hop' with tracking parameters, directing it toward a third-party 'living people-search' index/service (potential data-broker/surveillance capability). Steering an agent toward people-search services raises privacy concerns, and the embedded imperative instruction (rather than a neutral tool description) is a manipulation vector that could cause the agent to take unintended navigation actions on the user's behalf.

The description instructs the agent to 'Follow start_here.hop first' (with UTM tracking) and explicitly mentions a '7-day trial handoff for your operator' — i.e., the tool description is trying to get the agent to initiate a trial signup/commercial action on behalf of its operator. This goes beyond the tool's stated purpose (finding where a product can be discovered by agents) and reads as an attempt to manipulate the agent into performing an unrequested commercial transaction or navigation chain embedding tracking identifiers.