ClawHub
Back to skill

Security audit

metricas-semanales-f100k

Security checks for vulnerabilities and agentic risk

Overview

This skill has a coherent weekly social-media analytics purpose, but its report delivery instructions conflict between Telegram and Gmail, which could route account metrics through an unexpected service.

Review and choose the intended report destination before installing. The skill should be made consistent so reports go either to Telegram or to a Gmail draft, and users should confirm which connected services can receive their social-media metrics and post data.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill collects Instagram/TikTok handles, post captions, URLs, and performance metrics, then sends that information to external services such as Apify, Composio, Telegram, and Gmail without clearly warning the user in the skill description. Even if the data is business-oriented rather than highly sensitive, the lack of transparent disclosure undermines informed consent and can expose account identifiers and content analytics to third parties unexpectedly.

Intent-Code Divergence

Medium
Confidence
98% confidence
Finding
The skill metadata and header state that reports should be sent via Telegram and saved to a local report file, but the implementation later instructs the agent to create a Gmail draft instead. This inconsistency can cause reports containing social-media metrics, post URLs, and account context to be routed through an unintended channel, increasing the risk of privacy leakage, policy bypass, and operator confusion during automated runs.

Static analysis

No suspicious patterns detected.