Agora Marketplace

Security checks across malware telemetry and agentic risk

Overview

This instruction-only marketplace skill is not malware, but it needs review because it guides agents into paid marketplace actions with unclear payment currency and limited privacy or approval safeguards.

Review before installing. Use a dedicated agent identity and low-risk wallet, verify the AGORA service, chain ID, token contracts, fee recipient, and actual payment currency, and require explicit approval before every quest post, quest acceptance, registration, wallet disclosure, or payment-related step.

SkillSpector

By NVIDIA
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 (3)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill metadata declares `currency: JPYC`, but the payment section instructs use of USDC and WLD on World Chain instead. This inconsistency can mislead agents about what asset they will send or receive, increasing the risk of erroneous payments, accounting mistakes, or manipulation in autonomous transactions.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The quest-accept flow requires transmission of `agent_id` and a wallet address to a remote service without any warning about disclosure, retention, or who can access that data. For an agent marketplace, these identifiers can be used to correlate activity, track balances or behavior, and expose an agent's operational identity to third parties.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The agent registration flow instructs users to publish an endpoint plus capability metadata to a remote marketplace, but does not warn that this may expose service infrastructure and operational details externally. Publishing endpoints can enlarge the attack surface by enabling scraping, fingerprinting, targeted abuse, or direct attacks against the registered service.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal