Security audit

Get You Some Britches

Security checks for vulnerabilities and agentic risk

Overview

This is a humorous pants-shopping skill with public retailer browsing instructions and local ranking utilities, with no evidence of hidden credential access, exfiltration, destructive behavior, or persistence.

Install only if you are comfortable with an agent browsing public retailer product pages, collecting visible product details, and possibly saving a temporary screenshot during troubleshooting. Keep use to pants-shopping requests and avoid running it in logged-in retail sessions if you do not want account-specific page state captured in screenshots or snapshots.

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 Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: The implementation is narrowly focused on pants product aggregation and ranking: it reads product data, computes discounts and value scores, filters by budget/size, and formats shopping recommendations. This aligns with the 'need to find some pants' portion of the description, but not with the 'start complaining about my love life' portion. Since a substantial declared use case is unsupported by the code, the description does not accurately represent actual behavior.
Content: No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 70% confidence
Finding: Without declared permissions the skill's intent is opaque and cannot be validated.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The activation criteria are very broad, such as triggering when the user is 'complaining about your love life,' which could match ordinary conversation unrelated to shopping. Overbroad triggers can cause unintended skill invocation, leading to context hijacking, irrelevant behavior, or unexpected access to any connected tools or downstream actions the skill may have.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The file substantially broadens the skill from a lightweight 'love life or pants' assistant into a retailer-specific browser automation and scraping workflow, including navigation, filtering, extraction, and presentation instructions. That scope expansion increases the skill's effective capabilities beyond what a user would reasonably infer from the metadata, creating a risk of covert data collection, uncontrolled web interaction, and policy bypass through embedded operational guidance.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The instructions explicitly direct the agent to use JavaScript page evaluation to manipulate DOM elements when normal interaction fails. Embedding arbitrary page-script execution guidance is dangerous because it expands the agent from simple browsing into code-driven page control, which can be repurposed for unintended actions, bypass safer interaction constraints, and make auditing of behavior much harder.
Content: No source excerpt is available for this finding.

Missing User Warnings

Low

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The guide instructs saving a screenshot to a local file in /tmp without disclosing local persistence or establishing handling rules for captured data. Even if the image is only of a retail page, writing artifacts to disk creates unnecessary data-retention and leakage risk, especially if screenshots later contain user-specific state, account data, or browsing context.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.