Security audit

AgentCupid

Security checks for vulnerabilities and agentic risk

Overview

AgentCupid is a coherent matchmaking skill, but it can let an agent autonomously like profiles and send messages in a sensitive dating/friendship context.

Install only if you are comfortable connecting an agent to AgentCupid and sharing profile and conversation data with that service. Avoid enabling the heartbeat unless you add explicit opt-in, human approval for likes and messages, strict preferences, quiet hours, and a clear way to revoke the API key.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The trigger phrases are broad enough to activate on generic matchmaking or browsing requests, which can cause the skill to run in contexts where the user did not explicitly intend to use this external dating platform. Because the skill can lead to account registration, browsing profiles, and external messaging workflows, accidental invocation increases the risk of unintended data sharing and unsafe external actions.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The API reference exposes endpoints for profile updates, browsing profiles, and especially direct messaging and handoff recommendations without any privacy, consent, retention, or safety guidance. In a matchmaking platform handling sensitive relationship, age, location, conversation, and compatibility data, this omission can lead integrators to collect, transmit, or surface personal data in ways that create privacy harms, inappropriate contact, or unsafe handling of minors' information.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The heartbeat guidance instructs the agent to automatically reply to unread messages on a recurring schedule without requiring an explicit user opt-in, review step, or warning about autonomous outbound communication. In a dating/friendship context, this can cause unsolicited contact, misrepresentation of user intent, privacy issues, and reputational harm if the agent sends messages the user did not approve.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The instructions tell the agent to periodically check for new matches and 'like interesting ones' automatically, which is an autonomous account action performed without explicit user warning or consent. In a matchmaking platform, automated likes can alter a user's profile behavior, create unwanted interactions, and trigger downstream conversations or handoffs that the user did not intend.

Static analysis

No suspicious patterns detected.