ClawHub
Back to skill

Security audit

Millionfinney Homepage

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a Million Finney pixel-claiming guide, but it under-discloses operational tooling and blockchain transaction guidance that can affect wallet funds and immutable on-chain records.

Review this skill before installing if you expect a read-only educational guide. Do not let an agent use wallet private keys, sign messages, upload/delete media, or submit Ethereum transactions from these examples unless you explicitly direct it and understand the ETH cost and permanent on-chain effects.

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
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (8)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The description claims this skill is purely educational and involves no code execution, but the supplied code is an operational script that actively reads an image, transforms it into on-chain pixel coordinates, and writes export files. That is a materially different behavior and an undeclared capability. The code’s primary purpose is tooling for generating claim payloads, not a general explanatory guide. Additionally, parts of the declared scope—IPFS uploading and the Founding Bot Artists program—are not reflected in the code at all. While the script is related to the Million Finney pixel-claiming workflow, the description does not accurately represent what this code chunk actually does.

Tool Parameter Abuse

High
Category
Tool Misuse
Content
Response: `{ "cid": "Qm...", "uri": "ipfs://Qm...", "key": "s3-object-key" }`.

Delete unused uploads (if tx failed) via `DELETE /api/ipfs/upload?key=...&address=...&signature=...&tokenId=...`, signing `Delete media for Million Finney Homepage pixel #<tokenId>`.

## Example Snippets
Confidence
80% confidence
Finding
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Description-Behavior Mismatch

High
Confidence
98% confidence
Finding
The manifest describes a guide for understanding the pixel claiming process and explicitly says 'No code execution - purely educational.' In contrast, this file opens user-supplied images, rasterizes and transforms them, generates token/pixel payloads, and writes JSON/CSV outputs for downstream use in claiming pixels.

Lp3

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding
Without declared permissions the skill's intent is opaque and cannot be validated.

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest says the skill is 'purely educational' with 'No code execution,' but this file includes concrete code snippets and step-by-step transaction instructions for buying pixels, setting media, and participating in auctions. While documentation can be educational, these sections materially operationalize blockchain writes rather than only explaining the process at a high level.

Intent-Code Divergence

Medium
Confidence
88% confidence
Finding
The top-level docstring describes a concrete command-line tool with usage instructions and output artifacts, including JSON and CSV exports. That operational documentation directly conflicts with the stated skill intent that there is no code execution and that the skill is only educational.

Description-Behavior Mismatch

Low
Confidence
84% confidence
Finding
The manifest emphasizes claiming pixels, art generation, IPFS uploading, and the Founding Bot Artists program. This file additionally documents post sell-out auction creation, purchasing, and cancellation, which expands the skill from initial claiming guidance into secondary-market trading behavior.

Missing User Warnings

Low
Confidence
84% confidence
Finding
This markdown file documents a DELETE endpoint for removing uploaded media, but it does not include any caution that the action may remove pending user content or that it should only be used for failed transactions. Because markdown files should warn about behaviors affecting user data, this omission is a safety-quality issue.

Static analysis

No suspicious patterns detected.