OpenClaw Refactoring

• Outcome: Review
• Latest audit: May 28, 2026, 1:49 AM
• Version:
  1.1.0

Overview

This refactoring skill is mostly purpose-aligned, but it can rewrite many project files while its documentation overstates preview-by-default safety and implemented features.

Install only if you are comfortable with a tool that can modify project files. Run it inside version control, use --dry-run first, target a narrow directory, review the generated diff and .refactoring backups, and do not rely on advertised extract/inline/move/batch behavior without verifying support.

SkillSpector

By NVIDIA

Vulnerability Patterns

• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 88% confidence
• Finding: The skill documents file read/write refactoring behavior but does not declare corresponding permissions, which weakens security transparency and policy enforcement. In an agent environment, undeclared filesystem access can lead users or orchestration layers to grant trust under false assumptions, increasing the chance of unintended codebase modification or data exposure.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 91% confidence
• Finding: The documentation overstates implemented safety and refactoring capabilities while understating or omitting other behaviors, creating a trust mismatch between user expectations and actual execution. For a code-modifying skill, this can cause operators to invoke actions believing they are safe, supported, or limited in scope when they are not, leading to unintended source changes, failed refactors, or misuse of backup/analysis features.

Intent-Code Divergence

Medium

• Confidence: 94% confidence
• Finding: The documentation says modifications default to preview mode, but the examples show real write operations unless --dry-run is explicitly supplied. This safety inconsistency is dangerous because users may rely on the stated default and accidentally run destructive or large-scale code edits directly against a repository.

Intent-Code Divergence

Medium

• Confidence: 89% confidence
• Finding: Presenting extract, inline, move, and suggestion features as available while the roadmap says they are not yet implemented creates unsafe operational ambiguity. Users may attempt refactors that silently fail, partially modify files, or trigger fallback behavior, which is especially risky for a skill intended to automate source-code changes.

VirusTotal

VirusTotal findings are pending for this skill version.