Back to skill

Security audit

Cashback

Security checks for vulnerabilities and agentic risk

Overview

This cashback skill is mostly purpose-aligned, but its executable scripts allow broader inputs and execution paths than the stated three-merchant scope.

Review before installing. Use this only if you trust the publisher, the feima-lab API service, and the separately installed fx-base dependency. Avoid sending unsupported merchant links or URLs containing private tokens, and prefer a version-pinned, reviewed fx-base before using the API key.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
scripts/run.mjs:76
Finding

Path Traversal Enables Arbitrary JavaScript Execution and File Disclosure

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Error
Location
SKILL.md:14
Finding

Unpinned External Dependency Controls Authentication and Network Requests

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/link-convert.mjs:43
Finding

Documented Merchant Restrictions Are Not Enforced by Executable Entry Points

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (20)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill metadata explicitly restricts cashback support to Adidas, Space NK NL, and designwebstore DE, but the script help text and parameter handling advertise and accept arbitrary overseas merchant links such as Amazon, Nike, and iHerb. This mismatch can cause the agent to route unsupported merchants through the cashback conversion backend, violating declared scope and potentially enabling unauthorized affiliate processing, policy bypass, or data transmission to a third-party API for merchants the skill is not supposed to handle.

Content

No source excerpt is available for this finding.

Lp1

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'env' capability that is not listed in its permissions. This may indicate deceptive intent or missing permission declarations.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The script accepts any merchant name via --store and its help text explicitly lists arbitrary stores such as iHerb, Nike, and Amazon, which conflicts with the skill metadata that limits support to Adidas, Space NK NL, and designwebstore DE. This scope mismatch can cause the agent to route user queries and transmit merchant data to a backend for unsupported merchants, violating declared behavior and potentially enabling unauthorized data handling or policy bypass.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The README instructs users to run npx skills install ... without pinning a specific version of the installer or referenced package. This creates a supply-chain risk: users may fetch whatever version is current at install time, including a compromised or breaking release, and skill installation often executes package-managed code in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

A second unpinned npx skills install command appears for the cashback skill itself. As with the base dependency, this leaves installation subject to upstream package changes or takeover, which can result in arbitrary code execution during install or deployment of an unexpected skill version.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The README advertises personal cashback order lookup but does not clearly warn that order-related data may be retrieved from or transmitted to an external platform using the configured API key. In this context, the skill handles potentially sensitive purchase history, order status, amounts, and merchant metadata, so missing disclosure raises privacy and consent concerns and may cause users to share or query personal data unknowingly.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The activation guidance says users can 'directly send product links or ask merchant names, no commands needed' and includes broad phrases like cashback, 海淘返利, and 我的返利订单. Those triggers are loose enough that the skill may activate in conversations where the user did not intend cashback processing, increasing the chance of unintended link transformation, external API calls, or access to order data.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The skill instructs the agent to run npx skills install fangshan101-coder/fx-base if a dependency is missing, but does not pin a version or immutable source. This creates a supply-chain risk: a future malicious or compromised package version could be fetched and executed at runtime, especially since the dependency is required before API/auth instructions are read.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The title and instruction require output in Chinese (e.g. '输出模板' and '直接按模板填充输出'), and all examples are rendered in Chinese. This creates a language policy concern because the skill appears to force a specific language regardless of the user's preferred locale, with no opt-in or alternative documented.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The file title and instruction require the assistant to produce output in Chinese ('输出模板', '直接按模板填充输出') with no opt-in or alternative locale guidance. This is a natural-language policy concern because it imposes a specific language regardless of the user's preferred language.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The inline documentation contradicts the manifest by claiming support for broad merchants including Amazon, Nike, iHerb, and adidas, and by showing examples for unsupported stores. In an agent setting, misleading documentation is security-relevant because it can drive the model or operators to invoke out-of-scope actions, increasing the likelihood of policy violations and unintended disclosure of user-supplied shopping URLs to the external cashback API.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The natural-language description, help text, and user-facing messages are all written in Chinese, and the file does not indicate that the skill is China-specific or that users may choose another language. This can violate a language/locale policy when skills are expected to avoid forcing a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code presents usage guidance, examples, and data-flow disclosure entirely in Chinese, which imposes a specific language on users through natural-language content. The file does not provide an opt-in, alternative locale, or justification that the skill is intended only for a Chinese-speaking or region-specific audience.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The help text advertises cashback conversion and rebate lookup flows for merchants outside the manifest’s declared three-merchant scope. This can mislead users or higher-level agents into sending unsupported merchant URLs or queries to the backend service, creating scope bypass behavior and increasing the chance of unintended data handling or policy violations.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The inline error/help examples explicitly suggest Amazon and iHerb usage, which conflicts with the documented scope of the skill. In an agent setting, contradictory examples are dangerous because they can steer operators or automated tooling to invoke unsupported flows, causing out-of-scope requests to be processed or forwarded to the external API.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The inline metadata describes the tool as generically searching overseas merchants and querying rebate/commission plans, which materially overstates its permitted scope compared with the manifest's three-store cashback assistant. In an agent ecosystem, misleading embedded descriptions can influence tool selection and cause the skill to be invoked for unsupported tasks, increasing the chance of unintended external data transmission and policy noncompliance.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The natural-language description is written only in Chinese, and all user-facing help and error messages in this file are also Chinese, which effectively constrains the skill to a specific language without offering a user choice. The policy for SQP-3 asks to flag language or locale constraints when they are imposed without opt-in or documented justification.

Content

No source excerpt is available for this finding.