ClawHub
Back to skill

Security audit

Methodology Extractor

Security checks for vulnerabilities and agentic risk

Overview

This skill appears safe to install, but its packaged script only runs a demo and does not actually process the advertised paper input or output files.

Install only if a demo-level methodology comparison helper is acceptable. Do not expect it to batch-process real paper JSON files or write outputs until the --papers and --output paths are implemented or the documentation is corrected.

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
  • 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)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest describes a skill for extracting experimental methods from multiple papers for protocol comparison. However, the actual entrypoint ignores the declared --papers and --output arguments and only performs comparison on embedded demo data when --demo is used; otherwise it prints a message and exits. This is a semantic mismatch between the claimed operational capability and what the executable code actually does.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The argument help strings advertise a JSON file with paper data and an output JSON file, implying real batch extraction behavior. In practice, the program never reads args.papers or writes args.output, and the non-demo branch merely prints a usage hint. This documentation-level intent contradicts the implemented behavior.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.