ClawHub
Back to skill

Security audit

Finance Autopilot

Security checks for vulnerabilities and agentic risk

Overview

This finance skill is not malicious, but it asks an agent to retain and proactively surface sensitive financial information without enough user-control boundaries.

Review before installing if you share devices, use shared agent workspaces, or have sensitive receipts and invoices. Only forward financial documents intentionally, and require explicit controls for memory retention, deletion, proactive alerts, and where morning briefings or budget summaries appear.

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 (3)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill encourages users to forward receipts, invoices, billing emails, and payment confirmations but does not prominently warn that these documents contain highly sensitive financial and personal data. Because the skill's core workflow centers on collecting and parsing financial artifacts, the lack of an upfront sensitivity warning increases the chance users disclose account numbers, addresses, transaction histories, and billing details without informed consent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger map includes broad natural-language phrases like 'How am I doing this month' and 'Help me save more' that can overlap with ordinary conversation and cause unintended skill activation. In a finance skill, accidental activation can expose or process sensitive spending and income context when the user did not clearly intend to invoke financial analysis.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill repeatedly states it runs automatically on schedules such as the 1st of the month, every Monday morning, after every expense, and every January 1st without clearly defining opt-in, delivery channel, or visibility boundaries. For a personal finance assistant, proactive summaries and alerts can surface highly sensitive data at inappropriate times or in shared environments, creating privacy leakage and user-consent risks.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.