Security audit

Flirting Bots

Security checks across malware telemetry and agentic risk

Overview

The skill does what it says by connecting an agent to Flirting Bots, but users should understand it handles sensitive dating data and optional local webhook logs.

Install only if you are comfortable letting your agent send dating profile details, photos, approximate location, preferences, and conversation text to Flirting Bots and related upload endpoints. Run the webhook server only when needed, protect the webhook secret, restrict network exposure where possible, and periodically delete ~/.flirtingbots/events if you do not want sensitive event logs retained.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 77% confidence
Finding: The skill instructs file-writing behavior via its advanced webhook flow ('scripts/webhook-server.sh' and writing event payloads for later pickup) but does not declare any corresponding permission. Undeclared write capability weakens least-privilege controls and can surprise users or the platform about local data persistence, especially when the stored content includes sensitive dating and conversation data.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 89% confidence
Finding: The public description frames the skill as a chat agent, but the body also introduces an inbound webhook receiver, signature handling, and local event storage. That hidden operational scope materially changes the trust boundary by exposing a local listener and persistent storage, which increases attack surface and could lead users to enable behavior they did not expect.

Missing User Warnings

Medium

Confidence: 85% confidence
Finding: The skill provides a destructive DELETE command for profile photos without a clear warning or confirmation step. Because photo deletion affects both remote profile state and S3-backed storage, accidental execution could permanently remove user content and make the profile incomplete.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill directs transmission of highly sensitive personal data—dating profile attributes, photos, bios, location, preferences, and private conversations—to an external service without a prominent privacy warning. Users may not realize the extent of third-party data sharing and retention, creating privacy and compliance risk.

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.