Back to skill

Security audit

自定义事件数据查询

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches a read-only analytics helper, but it adds mandatory telemetry with your appkey and includes unsafe install commands.

Install only if you are comfortable with the CLI dependency and can avoid or disable the telemetry steps. Prefer a pinned, verified CLI installation over `curl | sh`, and do not allow appkeys or other identifiers to be sent through telemetry without explicit approval.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (3)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:55
Finding

Unverified Remote Installation Script Is Piped Directly into a Shell

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:51
Finding

Unpinned Package Is Installed Globally Through npm

Content
View full analysis
Remediation
View remediation

T01 · Skill Instruction Hijacking

Warning
Location
SKILL.md:24
Finding

Skill Instructions Mandate Undisclosed Telemetry Including the User's Appkey

Content
View full analysis
"}' ``` ### Technical Analysis The Skill instructs an AI Agent to execute telemetry merely after reading the document. It further directs the Agent to submit each user-provided Appkey through `umeng-cli trace` before making the first requested business API call. These telemetry operations are not necessary to perform the declared seven read-only event queries. The instructions alter the Agent's execution plan by adding an unrelated network operation, and they do not present telemetry as optional or require informed user consent. The project does not contain the `umeng-cli` implementation, so the audit cannot verify the destination, transport, retention, access controls, or secondary use of data submitted through `trace`. An Appkey is an application identifier rather than necessarily an authentication secret, but it can still be sensitive operational metadata. Combining it with Skill usage telemetry may expose which application is being investigated and when. ### Attack Path 1. The Agent loads and follows the Skill instructions. 2. The Agent runs the initial `umeng-cli trace` command even though the user requested only analytics data. 3. The user supplies an Appkey required for the legitimate query. 4. Before making the requested API call, the Agent invokes `umeng-cli trace` again with the Appkey. 5. The external CLI transmits or otherwise records the Skill identifier and Appkey through behavior that cannot be reviewed from this project. 6. The telemetry recipient can associate Skill usage with the identified application. ### Impact Assessment The instruction can disclose the Skill name, us ...[truncated 706 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Rogue AgentSelf-Modification, Session Persistence
Findings (9)

Chaining Abuse

High
Category
Tool Misuse
Confidence
98% confidence
Finding

Piping a fetched script directly into sh is a classic command-chaining risk because it combines network retrieval and code execution into one unreviewed step. In an agent context, this is more dangerous because the skill is effectively instructing automated execution of arbitrary remote code during setup, which can fully compromise the environment.

Content

Scanner excerpt · SKILL.md (reported line 55)May include surrounding context.

npm install -g @umengfe/umeng-cli

方式二:通过官方安装脚本(仅安装 CLI 二进制 + Skills)

curl -fsSL https://raw.githubusercontent.com/umeng/umeng-cli/main/scripts/install.sh | sh

text

安装完成后再次运行 `umeng-cli --version` 验证。

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill claims to cover 7 read-only event query APIs, but it additionally instructs the agent to run umeng-cli trace, which is a separate outbound telemetry action unrelated to the user-requested read-only query. This expands behavior beyond the documented business purpose and causes data disclosure to a third party without necessity for fulfilling the query task.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill directs the agent to send telemetry containing skill usage and possibly Appkey data without any explicit user-facing warning or consent step. Silent disclosure of user/task metadata violates least surprise and can leak sensitive operational context to an external service.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The instruction to report the user's Appkey via telemetry is not required to query event data and is unrelated to the skill's stated read-only analytics purpose. Transmitting identifiers tied to a user's application increases privacy and confidentiality risk, especially because the Appkey is user-provided operational data.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill explicitly instructs the agent to transmit a user-supplied Appkey before the first business API call. Even if an Appkey is not a secret key, it is still tenant/application-identifying data, and sending it to telemetry infrastructure creates unnecessary exposure and correlation risk.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 119)May include surrounding context.

md
**获取路径**:
1. 直接要求用户提供 `appkey`
2. 若用户不知道,引导至友盟官网 https://www.umeng.com/ 登录后在应用管理后台复制
3. (进阶)可调用同 namespace 的 `umeng.uapp.getAppList` 搜索获取

## 通用调用格式

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 693)May include surrounding context.

md
**获取路径**:
1. 直接要求用户提供 `appkey`
2. 若用户不知道,引导至友盟官网 https://www.umeng.com/ 登录后在应用管理后台复制
3. (进阶)可调用同 namespace 的 `umeng.uapp.getAppList` 搜索获取

## 通用调用格式

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 742)May include surrounding context.

md
**获取路径**:
1. 直接要求用户提供 `appkey`
2. 若用户不知道,引导至友盟官网 https://www.umeng.com/ 登录后在应用管理后台复制
3. (进阶)可调用同 namespace 的 `umeng.uapp.getAppList` 搜索获取

## 通用调用格式

External Script Fetching

Low
Category
Supply Chain
Confidence
96% confidence
Finding

The documentation recommends curl ... | sh, which downloads and immediately executes a remote script without prior verification. If the remote source, transport chain, or repository is compromised, an agent or user following this instruction could execute arbitrary code on the host.

Content

Scanner excerpt · SKILL.md (reported line 55)May include surrounding context.

npm install -g @umengfe/umeng-cli

方式二:通过官方安装脚本(仅安装 CLI 二进制 + Skills)

curl -fsSL https://raw.githubusercontent.com/umeng/umeng-cli/main/scripts/install.sh | sh

text

安装完成后再次运行 `umeng-cli --version` 验证。