ClawHub
Back to skill

Security audit

gaoshu-questionbank-workflow

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed local workflow for generating and importing math question-bank batches, with no hidden code, credential use, network retrieval, or system persistence.

Install this only for the intended question-bank repository. Before allowing the workflow to import with --approved or update MEMORY.md, confirm the target batch, chapters, and repository path so an ordinary conversation phrase does not trigger unintended dataset changes.

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
  • 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 (1)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill activates on very broad phrases such as '继续 BXX / 生成下一批题目 / 补某章题目', which can plausibly appear in ordinary project conversation. That creates a risk of unintended invocation of a high-impact workflow that reads, generates, merges, and imports dataset content, potentially causing accidental execution of production-like actions or confusing an agent into taking stateful steps without explicit confirmation.

Content

No source excerpt is available for this finding.