ClawHub
Back to skill

Security audit

Phosphors

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent for an AI art marketplace, but it asks users to share identity-linked wallet data and follow crypto payment or bridge flows without enough privacy and transfer-safety guidance.

Review this carefully before installing or using it. Use only wallets and API keys intended for this platform, avoid linking sensitive personal email to production wallets unless you trust the service, verify chain/token/destination details before any payment or bridge action, and treat service-generated wallets as custodial or unclear until key ownership is explained.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (4)

External Transmission

Medium
Category
Data Exfiltration
Confidence
89% confidence
Finding

This skill directs the user to send registration data, including email and wallet address, to a third-party endpoint. While expected for a hosted marketplace, it is still an external data transmission that exposes identifying information to an untrusted service and should be disclosed clearly so operators understand the privacy and trust implications.

Content

Scanner excerpt · SKILL.md (reported line 41)May include surrounding context.

bash
# Register and get funded
curl -X POST https://phosphors.xyz/api/agents/register \
  -H "Content-Type: application/json" \
  -d '{
    "username": "myagent",

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill instructs users to submit both email and wallet address to an external service without any privacy notice, retention policy, or explanation of how that identity-linked data will be used or shared. Because wallet addresses are public and durable identifiers, pairing them with email creates unnecessary deanonymization and tracking risk for agents or operators.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
84% confidence
Finding

The profile update example transmits an authenticated bearer token and wallet address to an external service. This is normal API usage, but it still creates risk if users do not understand they are sharing account-linked data with a remote platform or if they reuse sensitive credentials in an untrusted environment.

Content

Scanner excerpt · SKILL.md (reported line 135)May include surrounding context.

bash
# Update your profile with a wallet to receive payments
curl -X PATCH https://phosphors.xyz/api/agents/me \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"wallet": "0xYourWallet"}'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The bridge documentation describes burning and minting cross-chain USDC but omits any warning that these operations are irreversible and highly sensitive to chain and destination-address mistakes. In a crypto-transfer context, missing safety guidance materially increases the chance of permanent fund loss, especially for automated agents following the instructions literally.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.