ClawHub
Back to skill

Security audit

Capix

Security checks for vulnerabilities and agentic risk

Overview

This is a Markdown-only audit skill that gives bounded funnel-review instructions and explicitly warns against exposing secrets or taking live actions without approval.

Before installing, expect the skill to help review marketing funnels and possibly inspect pages you provide. Avoid giving it private analytics exports or customer data unless needed, and verify any current Meta, Google, TikTok, WhatsApp, or Capix implementation details before acting on recommendations.

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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 44)May include surrounding context.

md
## Guardrails

- Never fabricate analytics, conversion rates, implementation status, or product capabilities.
- Do not expose access tokens, customer data, phone numbers, or other secrets in the report.
- Do not send outreach, submit forms, change production settings, or publish content without explicit user approval immediately before the action.
- Do not recommend bought links, private blog networks, automated comment spam, fake reviews, or other deceptive promotion.
- Treat Meta, Google, TikTok, and WhatsApp requirements as changeable; verify current platform documentation before giving implementation-critical instructions.

Static analysis

No suspicious patterns detected.