Security audit

Near Airdrop Hunter

Security checks across malware telemetry and agentic risk

Overview

The skill does not appear to steal data or submit blockchain transactions, but its airdrop “claim” flow may make users think a claim succeeded when it only shows a link and records local tracking.

Install only if you understand this as a manual airdrop link and tracking helper, not an automated verifier or claimant. Independently verify every airdrop URL before connecting a wallet, review any wallet signature or approval carefully, and do not treat the local “claimed” record as proof that rewards were actually claimed.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 92% confidence
Finding: The skill documentation promises discovery, eligibility checks, and claiming, but the described/observed behavior includes local state writes and simulated tracking without clearly matching those claims. This is dangerous because users may believe eligibility was verified or rewards were actually claimed, leading to unsafe wallet actions, bad operational decisions, or trust in misleading outputs.

Description-Behavior Mismatch

Medium

Confidence: 97% confidence
Finding: The skill metadata and CLI verbs imply that it can actively check eligibility and claim rewards, but the implementation only prints URLs and instructions. This mismatch can mislead users or higher-level agents into believing a security-sensitive action was performed when it was not, creating unsafe automation assumptions and inaccurate state tracking.

Intent-Code Divergence

Medium

Confidence: 98% confidence
Finding: The 'claim' command strongly suggests an actual reward claim, but it only logs a manual URL and writes a local 'claimed' tracking entry. In an agent skill context, this is dangerous because downstream tooling or users may treat the claim as completed, causing missed rewards, duplicate workflows, or incorrect audit/state records.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The README instructs users to run a claim command for an airdrop without clearly warning that claiming may trigger wallet interactions, transaction signing, fees, or other irreversible asset-affecting actions. In a crypto skill focused on discovering and claiming rewards, this omission is more dangerous because users may assume the operation is low-risk and execute it without verifying contract legitimacy or understanding wallet consequences.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The skill describes wallet-affecting actions like claiming airdrops and also stores tracking data locally, but it does not prominently warn users about transaction implications, external trust requirements, or local persistence. In the context of crypto airdrops, this is especially risky because users may connect wallets or act on claim prompts without understanding that third-party links, token approvals, or persistent local records may be involved.

Missing User Warnings

Low

Confidence: 85% confidence
Finding: The script persists account identifiers and claim activity to a file in the user's home directory without clear consent, disclosure, retention policy, or permission hardening. While the data is not highly sensitive by itself, wallet/account activity can still expose user behavior and create privacy issues on shared systems.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.