Security audit

Subscription Killer

Security checks for vulnerabilities and agentic risk

Overview

The skill locally analyzes a user-provided bank CSV for recurring subscriptions and does not show hidden network, persistence, credential, or file-writing behavior.

This skill appears safe to install for local CSV analysis, but bank transaction data is sensitive. Review results before acting on cancellation advice, especially where refunds, credits, or mixed debit/credit sign conventions may be present.

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
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
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)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 70% confidence
Finding: Without declared permissions the skill's intent is opaque and cannot be validated.

Intent-Code Divergence

Medium

Confidence: 98% confidence
Finding: The implementation contradicts its own security/data-handling logic by converting any non-zero amount to abs(val), which includes credits and refunds instead of excluding them. In a finance-analysis skill, this can silently misclassify incoming transactions as subscription debits, producing false recurring charges, misleading cancellation advice, and privacy-sensitive downstream decisions based on incorrect financial interpretation.

Vague Triggers

Low

Confidence: 80% confidence
Finding: This JSON file is a manifest-scoped file type for SQP-1, but it contains only service names and cancellation URLs with no indication of how or when the skill should activate. If this file is used as part of skill routing or invocation metadata, the absence of explicit trigger constraints or negative examples could contribute to overly broad matching.

Static analysis

No suspicious patterns detected.