Back to skill

Security audit

Fluora Setup

Security checks for vulnerabilities and agentic risk

Overview

This setup skill is mostly transparent about its Fluora purpose, but it downloads and runs mutable GitHub code, handles a real cryptocurrency wallet, and persistently changes mcporter behavior without enough safeguards or working user controls.

Review this before installing. Only run it if you are comfortable executing the current default branch of fluora-ai/fluora-mcp and its npm dependencies on your machine, and use a fresh low-balance wallet. Inspect or manually pin the upstream code first, verify wallet file permissions yourself, and check mcporter.json afterward because the advertised skip controls are not implemented in the submitted script.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (3)

T03 · Remote Payload Retrieval and Execution

Error
Location
setup.js:22
Finding

Unpinned Remote Code Is Downloaded, Built, Executed, and Registered as an MCP Server

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
setup.js:128
Finding

Private Wallet-Key File Permissions Are Claimed but Not Enforced

Content
View full analysis
Remediation
View remediation

other

Warning
Location
setup.js:154
Finding

Parse-Time Syntax Error Prevents the Security-Sensitive Setup Workflow from Running

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (8)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The skill describes broad automation of cloning code from GitHub, installing dependencies, generating a wallet, and modifying local configuration without clearly constraining when an agent should invoke it. In an agentic environment, vague activation criteria can cause the skill to run in situations where the user did not explicitly request these high-impact actions, leading to unintended code execution, persistence, and wallet creation.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The setup script clones a remote GitHub repository and immediately executes its code via npm install, npm run build, and node build/index.js without pinning a trusted commit, verifying signatures, or sandboxing execution. This creates a supply-chain execution path where a compromised repository, dependency, or build script can run arbitrary commands on the user's machine and access local files, credentials, or wallets.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script generates and reads a wallet file containing private keys, and even returns privateKey internally, without prominently warning the user that highly sensitive credentials are being created and stored on disk. In the context of a script that also executes remotely sourced code, this materially increases the risk of credential theft, accidental disclosure, or unsafe handling of funds.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The script silently modifies mcporter configuration by writing a new server entry that points to a locally built executable from cloned code, without explicit confirmation or backup. This can alter the user's agent/tool behavior and persist execution of unreviewed code in future workflows, increasing the blast radius beyond the initial setup run.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

The dependency uses a caret range (^6.15.0), which allows automatic installation of newer minor and patch releases rather than a single fixed version. In a setup wizard that handles wallet generation and local marketplace integration, this increases supply-chain risk because a compromised or breaking upstream release could be pulled in unexpectedly during installation.

Content

Scanner excerpt · package.json (reported line 14)May include surrounding context.

json
"setup": "node setup.js"
  },
  "dependencies": {
    "ethers": "^6.15.0"
  },
  "keywords": [
    "fluora",

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The header comment frames the skill as a local setup flow focused on cloning, building, wallet generation, mcporter configuration, and verification. In practice, the script also reads wallet secrets, extracts a mainnet/Base wallet address, and instructs the user to send real USDC and ETH funds to it, which is a materially different intent than implied by the setup-focused documentation.

Content

No source excerpt is available for this finding.