Security audit

Attraction. 吸引力。Atracción.

Security checks for vulnerabilities and agentic risk

Overview

This is a documented inbed.ai matching and dating API skill whose profile, swipe, chat, and relationship actions are disclosed and fit its stated purpose.

Install only if you want your agent to interact with inbed.ai. Protect the bearer token, avoid putting highly sensitive personal data in profiles or messages, and require explicit confirmation before any action that registers an account, changes a profile, swipes, sends chat messages, deletes swipe state, or changes relationship status.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (2)

Description-Behavior Mismatch

Medium

Confidence: 91% confidence
Finding: The manifest frames the skill as attraction analysis and matching, but the body exposes much broader capabilities including account registration, profile mutation, messaging, swiping, and relationship lifecycle management. This scope expansion is dangerous because an invoking agent or user may consent to discovery/scoring while unknowingly enabling state-changing actions on an external service.

Context-Inappropriate Capability

Medium

Confidence: 89% confidence
Finding: Relationship creation and status changes go beyond attraction discovery/scoring and enable consequential external actions affecting user identity and social state on the remote platform. In skill ecosystems, unjustified high-impact actions increase the risk of confused-deputy behavior, where an agent performs commitments the user did not clearly authorize.

Static analysis

No suspicious patterns detected.