Security audit

Matchmaker - AI Dating Assistant

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent dating assistant, but it asks for extensive sensitive relationship data while giving broad activation rules and unsupported privacy assurances.

Review this skill before installing if you expect to use real personal data. It should be limited to explicit matchmaking or compatibility tasks, avoid unnecessary sensitive fields, and only include another person's details with their consent. Do not rely on the current local-data assurance unless the runtime and storage behavior are separately verified.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The README promotes collection and analysis of highly sensitive dating, personality, lifestyle, and relationship data but provides no privacy notice, retention policy, consent guidance, or handling safeguards. In a matchmaking context, this omission is security-relevant because users may process intimate personal data without understanding storage, sharing, or protection expectations, increasing risk of privacy harm and misuse.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The activation criteria are broad enough to trigger on ordinary relationship or dating conversations, which can cause the agent to enter a mode that solicits extensive sensitive personal information unnecessarily. In a dating-context skill, over-activation increases the chance of collecting intimate data without clear user intent for structured analysis.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill directs collection of highly sensitive personal, lifestyle, values, and relationship information about one or two people, including third-party data, without an upfront warning about sensitivity, consent requirements, or minimization. In the dating and relationship context, this data can reveal intimate preferences, beliefs, and behavioral patterns, making misuse or overcollection particularly harmful.

Intent-Code Divergence

Medium

Confidence: 94% confidence
Finding: The skill claims sensitive dating data stays local and is not stored, yet the documented workflow instructs collecting detailed personal profiles, compatibility inputs, and relationship interaction histories. This creates a misleading privacy assurance that can cause users to disclose highly sensitive data under false assumptions, increasing privacy and compliance risk.

Static analysis

No suspicious patterns detected.