ClawHub
Back to skill

Security audit

RumbleTipAI

Security checks for vulnerabilities and agentic risk

Overview

The skill is transparent about being an autonomous cryptocurrency tipping agent, but it asks for real wallet/payment authority with under-scoped natural-language controls and unpinned wallet dependencies.

Review this carefully before installing. Only use it with small spending limits, confirm exactly how wallet seed storage and transaction approval work in the actual extension, and prefer a build with pinned dependency versions or a lockfile. Treat all autonomous tips as real, irreversible cryptocurrency payments.

A.I.G

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Error
Location
SKILL.md:17
Finding

Unpinned Wallet Dependencies Create a Supply-Chain Risk

Content
View full analysis
Remediation
View remediation

SkillSpector

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 (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill exposes broad natural-language invocations such as deleting rules, showing wallet data, or changing budgets without documenting strict trigger constraints, confirmation boundaries, or authorization checks at the invocation layer. In an autonomous financial agent, ambiguous everyday phrases materially increase the risk of unintended high-impact actions being triggered by casual conversation, prompt injection from page content, or misinterpretation by the LLM.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% confidence
Finding

The skill intentionally sends operational context to external services, including OpenAI for decision-making and blockchain/RPC infrastructure for payment execution. While this is part of the product design, it still creates a real security and privacy exposure because financial metadata, creator associations, wallet recipient data, and behavioral data like watch time leave the local environment and may be logged, correlated, or mishandled by third parties.

Content

Scanner excerpt · SKILL.md (reported line 210)May include surrounding context.

md
| Endpoint | Purpose | Data Sent |
|----------|---------|-----------|
| `https://api.openai.com/v1/chat/completions` | AI reasoning for tip decisions and chat assistant | Creator name, watch time, rule config, budget state (no PII) |
| `https://polygon-rpc.com` / `https://arb1.arbitrum.io/rpc` / `https://eth.drpc.org` | Blockchain RPC for on-chain transfers | Transaction data (recipient address, amount, token contract) |
| `https://rumble.com` (HTMX endpoints) | Extract creator wallet address from Rumble's native tip button | Page-local HTMX requests only (no auth tokens sent) |
| `wss://electrum.blockstream.info:50004` | Bitcoin network via Electrum WebSocket | BTC transaction data |