Back to skill

Security audit

coding

Security checks for vulnerabilities and agentic risk

Overview

This skill is read-only and purpose-aligned enough to install, though it includes an unreferenced linter script that does not match the main description cleanly.

Before installing, understand that the main skill promises to read only a style reference file you name. The bundled linter script is separate, currently not described by the main skill, and even fails to run because of an argparse typo; treat it as nonessential unless the publisher fixes and documents it.

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
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The declared purpose says this should be used when a user asks the agent to follow coding conventions written in the user's reference file (for example, 'use the style in file X' or 'follow my style guide'). The actual code does not inspect any reference file at all; in fact, its docstring explicitly states it does not read, write, or keep any reference file. Its primary function is generic linting of a Python file against hardcoded defaults and simple command-line overrides. That is a materially different purpose from deriving conventions from a user-supplied file, so this is a clear description/behavior mismatch.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The docstring states that the linter 'does not read ... any ... reference file,' but the implementation opens and reads the path provided by the user in lint_file(). While this may have intended to distinguish against a separate style reference file, the wording contradicts the actual behavior of reading a file from disk.

Content

No source excerpt is available for this finding.