ClawHub
Back to skill

Security audit

AI-Friendly Structured Writer

Security checks for vulnerabilities and agentic risk

Overview

This is a content-formatting skill that can read and optionally write user-selected files, with no hidden network access, credential use, persistence, or destructive behavior.

Review generated output before publishing because the included script is basic and may not perform all advertised restructuring. Use explicit input and output paths, and avoid running it on sensitive drafts unless you are comfortable with the selected agent reading that content.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (3)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The documented behavior materially overpromises capabilities that are not actually implemented, creating a trust gap between what operators expect and what the skill does. In security terms, this can mislead users and orchestrators into relying on transformations, safeguards, or content handling that never occur, which is especially risky when the skill may write outputs or be chained into publishing workflows.

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill references a file-writing capability via the documented restructuring tool and output workflow, but it does not declare any explicit tool scope or permissions boundaries. In an agent environment, undeclared write capability increases the risk of unintended file modification, misuse by downstream orchestration, or reviewers underestimating what the skill can change.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger language is broad enough to activate on many ordinary writing or formatting requests, causing the skill to run outside narrowly intended scenarios. Over-broad activation expands attack surface and can lead to unnecessary file modifications, unintended content rewriting, or invocation in contexts where GEO-style formatting is inappropriate.

Static analysis

No suspicious patterns detected.