Security audit

BLE → $ANIMA Minter

Security checks across malware telemetry and agentic risk

Overview

The skill does what it advertises, but it handles nearby Bluetooth device identifiers that users should treat as privacy-sensitive.

Install only if you intentionally want to scan nearby BLE devices. Treat terminal output and anima_dag.gpickle as sensitive because they can reveal or correlate nearby devices; avoid sharing them and delete them when no longer needed. Consider pinning dependency versions before installation.

SkillSpector

By NVIDIA

Vulnerability Patterns

Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (7)

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The code actively scans nearby BLE devices and collects their MAC addresses, which are device identifiers and can enable tracking of people or assets in physical proximity. In a skill with no clear user-facing purpose, consent flow, or minimization controls, this is an unjustified sensitive data collection capability and creates privacy and surveillance risk.

Context-Inappropriate Capability

Medium

Confidence: 94% confidence
Finding: The skill persists collected device-derived identifiers to a local gpickle graph file, creating a durable record of nearby devices over time. Even though the MACs are hashed, the hashing uses a hardcoded static salt, so the identifiers remain linkable across runs and can still support tracking or correlation if the salt becomes known.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill advertises scanning nearby BLE MAC addresses, hashing them, storing them in a local DAG, and optionally gossip-syncing derived data, but it does not warn users that this collects identifiers from nearby devices and may propagate related data to other nodes. Even if MACs are hashed, they originate from third-party device identifiers and can still create privacy, consent, and tracking risks, especially when combined with salting, persistence, and distributed sharing.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The BLE callback prints detected MAC addresses and their hashes to output without any warning, notice, or consent mechanism. This exposes sensitive nearby-device information to logs or consoles and increases the chance of unauthorized observation or secondary collection by anyone with access to output streams.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The code writes collected scan results to anima_dag.gpickle without any prior disclosure to the user that sensitive device-derived data will be retained. Silent persistence increases privacy risk because users cannot make an informed decision, and the stored graph can later be reused for correlation or surveillance purposes.

Unpinned Dependencies

Low

Category: Supply Chain
Content: aioblescan networkx
Confidence: 95% confidence
Finding: aioblescan

Unpinned Dependencies

Low

Category: Supply Chain
Content: aioblescan networkx
Confidence: 95% confidence
Finding: networkx

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.