ClawHub
Back to skill

Security audit

Security Audit

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward Move smart contract audit checklist skill with no hidden execution, persistence, or credential use.

Installers should understand this skill may trigger on broad security wording, so it is best suited for Aptos Move contract review contexts. It does not request secrets and explicitly tells the agent not to read private-key config or .env files.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are broad, generic, and likely to match ordinary user requests such as 'is this secure' or 'security check,' causing unintended invocation. In an automated agent environment, this can lead to the skill activating in contexts the user did not explicitly intend, increasing the chance of prompt-scope confusion, unexpected actions, or disclosure of analyzed content to the skill pipeline.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 59)May include surrounding context.

md
// Safe to proceed
}

// ❌ WRONG: No verification
public entry fun update_config(admin: &signer, value: u64) acquires Config {
    let config = borrow_global_mut<Config>(@my_addr);
    config.value = value; // Anyone can call!

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description says the skill audits using a '7-category checklist'. However, the body defines Step 1 as 'Run Security Checklist' and then enumerates seven categories, followed later by a separate 'Step 8: Testing Audit', making the documented workflow eight steps/categories in practice. This is a direct contradiction between top-level documentation and the actual documented procedure.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.