Back to skill

Security audit

电商平台结算核对(免费版)

Security checks for vulnerabilities and agentic risk

Overview

The bundled checker is local and coherent, but its optional upgrade instructions ask an AI agent to run an unpinned external installer and continue into a purchase flow.

Install the free checker only if you are comfortable with a local Chinese-language settlement-audit tool. Do not copy the optional upgrade prompt to an AI assistant or run the `npx ... @latest` installer unless you separately verify the package source, version, and payment flow and approve each step yourself.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:63
Finding
Mutable Third-Party Package Execution in Optional Upgrade Workflow## Vulnerability Details **File Location**: `SKILL.md`, lines 63-70 **Vulnerability Type**: Unsafe execution of an unpinned third-party dependency **Risk Level**: Medium ### Vulnerable Code ```text Copy the following text verbatim to your AI assistant, and it will guide you through installation and purchase (wallet authorization or payment must be confirmed by you): Please purchase and install the Skill “Platform Settlement Check · Buyout Edition” through SkillPay. Official installation and purchase guide: https://skillpay.alipay.com/buildguide.md First read the official guide, verify the source, and complete the first step in the guide: npx -y @alipay/agent-payment@latest install After installation, read the SKILL.md of the newly installed skillpay-onboarding, then follow its instructions to purchase and install product S0806000206825371 from merchant 2088702818190040. I authorize dependency installation and initiation of the purchase; when wallet authorization or payment is needed, provide the official link or QR code for my confirmation and do not confirm it on my behalf. After confirmation, continue product installation and verification. ``` ### Technical Analysis The optional commercial-upgrade instructions direct an AI agent to execute: ```bash npx -y @alipay/agent-payment@latest install ``` The `@latest` selector is mutable and does not identify a specific reviewed package artifact. The `-y` option automatically accepts the package execution prompt. No pinned version, package lock, integrity hash, vendored source, or signature-verification procedure is provided. As a result, the code executed when a user follows these instructions may differ from the code that existed when this Skill was audited. The workflow also delegates subsequent actions to the `SKILL.md` file of the newly installed package, creating a secondary trust boundary in which remotely supplied instructions can influence later installation and pu ...[truncated 1770 chars]
Remediation
## Remediation Suggestions 1. Remove executable installation instructions from the local settlement-checking Skill and place the commercial upgrade process in separate, clearly identified documentation. 2. Replace `@latest` with an explicitly audited package version. 3. Verify the downloaded package against a publisher-provided cryptographic integrity hash or signature before execution. 4. Avoid automatic confirmation through `npx -y`; display the exact package name, version, source, and requested actions before obtaining explicit user approval. 5. Review and pin every transitive dependency through a lockfile or equivalent reproducible installation mechanism. 6. Do not automatically trust or execute instructions from a newly installed `SKILL.md`. Subject the installed Skill and its scripts to an independent security review before granting tools or following further instructions. 7. Require separate, explicit user authorization for dependency installation, network access, product installation, wallet interaction, and payment initiation. 8. Run any unavoidable installer in a restricted sandbox with minimal filesystem access, no ambient credentials, no elevated privileges, and narrowly controlled network access.
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 (9)

Ae1

High
Category
analysis-evasion
Content
node scripts/run.mjs --sample # 先看样例
Confidence
100% confidence
Finding
Referenced artifact was not completely inspected

Ae1

High
Category
analysis-evasion
Content
node scripts/run.mjs --sample # 先看样例
Confidence
100% confidence
Finding
Referenced artifact was not completely inspected

Natural-Language Policy Violations

Medium
Confidence
88% confidence
Finding
The skill's primary display names, summaries, descriptions, usage instructions, and purchasing prompt are all written in Chinese, while the only English description is partial and still references Chinese terminology. There is no statement that the skill is Chinese-only, no justified locale restriction, and no option for users to choose language, which can violate language/locale policy requirements.

Ssd 4

Medium
Confidence
98% confidence
Finding
The skill contains an embedded prompt that tells an AI assistant to install another skill, read that skill's instructions, and continue through a payment/purchase flow. This is dangerous because it attempts to delegate trust transitively to newly fetched instructions and external content, enabling prompt-injection-style tool use, unauthorized workflow chaining, and potentially unsafe purchases or installs.

Rp1

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding
The skill embeds a command that installs and executes an external package via `npx -y @alipay/agent-payment@latest`, which is effectively unpinned and pulls whatever the registry serves at execution time. This creates a supply-chain and remote-code-execution risk, especially because the surrounding text instructs an AI assistant to run it as part of a purchase/install workflow.

Natural-Language Policy Violations

Medium
Confidence
92% confidence
Finding
This code file contains user-facing natural-language strings and documentation exclusively in Chinese, including the contract, scope, and guidance messages. Because the file does not offer user opt-in for language/locale or explain that the skill is intentionally limited to a Chinese-language/regional context, it may violate language/locale policy.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
This code file contains extensive natural-language instructions, help text, and runtime messages exclusively in Chinese. That effectively forces a specific language/locale for all users without offering an alternative or opt-in, which matches the language/locale policy violation criteria.

Intent-Code Divergence

Low
Confidence
95% confidence
Finding
L077 的注释称非 JSON 输入可为“直接粘贴的合同全文”,但该脚本从清单描述、帮助文本到错误提示均表明其用途是平台结算单核对。代码实际只是把非 JSON 内容当作原始文本交给结算单核对引擎,这与“合同全文”这一文档表述形成明显意图层面的误导。

Natural-Language Policy Violations

Low
Confidence
89% confidence
Finding
The JSON value consists entirely of Chinese headers and labels, which indicates a fixed language/locale in the skill content. Under the policy rule, forcing a specific language without user opt-in or a documented regional justification can be a natural-language policy violation.

Static analysis

No suspicious patterns detected.