ClawHub
Back to skill

Security audit

Safeflow Sui Skill

Security checks across malware telemetry and agentic risk

Overview

Review recommended: the skill is purpose-built for SafeFlow/Sui payment testing, but it stores delegated payment capability data and can run a real payment test flow without strong confirmation or fully reviewed helper code.

Install only if you intentionally want SafeFlow/Sui payment-test automation. Use testnet or tightly capped funds, inspect the surrounding agent_scripts and package dependencies before running the E2E flow, verify recipients and amounts, keep generated .safeflow files private and out of version control, and revoke or rotate session capabilities after testing.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill directs the operator to save `walletId` and `sessionCapId`, which are sensitive runtime identifiers tied to autonomous payment capability, without warning about secure storage, retention, redaction, or exposure risks. In this context, those values may enable unauthorized payment actions or session abuse if logged, persisted insecurely, or copied into chat history and scripts.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The document directs operators to save owner-provided wallet and session identifiers into local config and environment artifacts, but it does not warn that these values may be sensitive operational secrets or capability references. In this skill context, those artifacts are later used for autonomous payment execution, so insecure storage, accidental git commits, shell history exposure, or overly broad filesystem permissions could enable unauthorized use or replay of payment-related capabilities.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The document presents a 'One Command E2E' flow that creates an intent, executes it once, and reports a transaction digest and Walrus blob ID, but it does not clearly warn that these steps can trigger real on-chain payment activity and evidence uploads. In this skill context, operators are handling real wallet/session identifiers and a real Publish API base URL, so a user could reasonably run the command as a harmless test and unintentionally spend funds or publish data.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
This script creates a real payment intent and then runs the consumer flow that may execute blockchain-backed side effects, but it does not present a clear interactive warning, dry-run default, or explicit confirmation gate before doing so. In the context of a SafeFlow/Sui payment skill, that omission is materially risky because operators may assume this is a harmless test while it can trigger actual fund movement or on-chain actions.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.