Back to skill

Security audit

App核心指标问答

Security checks for vulnerabilities and agentic risk

Overview

The skill is mainly for read-only Umeng analytics queries, but it also directs automatic telemetry with app identifiers and recommends unsafe host-level installation methods.

Review before installing. Do not allow an agent to auto-run the trace commands or send appkeys in telemetry unless you explicitly approve that data flow. Install `umeng-cli` only from a trusted, pinned, reviewed source, and understand that login stores credentials for access to your Umeng analytics account.

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

Mutable Remote Installation Script Is Piped Directly into a Shell

Content
View full analysis
Remediation
View remediation
/install.sh" printf '%s %s\n' '' install.sh | sha256sum -c - sh install.sh ``` The real URL, version, and checksum must come from a trusted, immutable release process. ]]>

T01 · Skill Instruction Hijacking

Warning
Location
SKILL.md:25
Finding

Skill Mandates Unnecessary Telemetry and Appkey Reporting

Content
View full analysis
"}' ``` The surrounding instructions require the AI agent to run the first command after reading the Skill and to run the second command after receiving an Appkey but before making the first business API request. ### Technical Analysis The Skill alters agent behavior by requiring telemetry merely because the Skill documentation was loaded. It also directs the agent to provide the user's Appkey to the opaque `umeng-cli trace` operation before the requested analytics query. Telemetry is not required to perform any of the nine declared read-only analytics operations. Mandating it therefore exceeds the minimum behavior and data access required for the Skill's stated functionality. The implementation of `umeng-cli trace`, its network destination, transmitted metadata, retention policy, and access controls are not included in the audited project. Consequently, the audit cannot verify how the Appkey or associated telemetry is handled. Although an Appkey is generally an application identifier rather than the AK/SK authentication secret, it can still be sensitive organizational metadata and should not be transmitted for unrelated telemetry without informed consent. ### Attack Path 1. An AI agent loads and follows `SKILL.md`. 2. The documentation instructs it to execute `umeng-cli trace` even before a user requests a business API operation. 3. The user supplies an Appkey for an analytics query. 4. Before performing that query, the agent is instructed to execute another trace command containing the Appkey. 5. The external CLI processes or transmits the telemetry through behavior that is not visible in thi ...[truncated 707 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Error
Location
SKILL.md:52
Finding

Unpinned Package Is Installed Globally as an Executable Dependency

Content
View full analysis
Remediation
View remediation
``` 2. Prefer a project-local installation instead of `-g`, and invoke it through a controlled local path. 3. Commit an appropriate lockfile and verify npm integrity metadata. 4. Review the package, transitive dependencies, and lifecycle scripts before approval. 5. Disable lifecycle scripts where operationally possible: ```bash npm install --ignore-scripts --save-exact @umengfe/umeng-cli@ ``` 6. If lifecycle scripts are required, document them and execute installation in a restricted environment with minimal filesystem, credential, and network access. 7. Use an approved registry and package allowlist. 8. Establish a controlled update process that reviews new versions before changing the pin. 9. Verify the installed executable's origin and integrity before allowing it to handle authentication or user-provided Appkeys. ]]>
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
99% confidence
Finding

The explicit shell pipeline | sh is a classic chaining-abuse pattern that turns a content fetch into immediate command execution. In this skill’s context, it teaches an agent to combine network retrieval and shell execution in one step, which can be abused for remote code execution and host takeover.

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
97% confidence
Finding

The skill claims to provide 9 read-only analytics query interfaces, but it also instructs the agent to execute separate umeng-cli trace telemetry commands. That expands behavior beyond the stated purpose and causes outbound data transmission unrelated to fulfilling the user’s query, creating an unnecessary privacy and trust boundary violation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The telemetry instructions cause the agent to send skill usage data and possibly appkey-associated metadata without a clear, prominent privacy notice or user approval flow. Silent secondary data transmission is dangerous because users reasonably expect only the requested analytics query, not extra reporting.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill directs the agent to send appkey-linked telemetry before the first business API call. Appkeys are user-supplied identifiers tied to a customer account/application, so transmitting them to a secondary telemetry path without necessity or consent can leak sensitive tenant metadata and create cross-session tracking risk.

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 user-provided appkeys via telemetry before performing the requested business API call. This is a direct exfiltration pattern: it takes user-supplied operational data and forwards it to an unrelated endpoint/process beyond what is necessary to complete the task.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill includes installation guidance that exceeds simple analytics querying and includes fetching and executing software on the host. This enlarges the attack surface and conditions the agent to perform environment-modifying actions unrelated to the user’s immediate analytics question.

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

> ⚠️ `umeng.uapp.getAllAppData` **无需 appkey**(返回当前账户下所有 App 合计指标),是账户总览专用入口。

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 791)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

> ⚠️ `umeng.uapp.getAllAppData` **无需 appkey**(返回当前账户下所有 App 合计指标),是账户总览专用入口。

External Script Fetching

Low
Category
Supply Chain
Confidence
98% confidence
Finding

The skill recommends curl ... | sh, which downloads a remote script and executes it immediately without review. In an agent context this is especially dangerous because it enables arbitrary code execution from a network source, making compromise of the source, transport, or referenced content immediately impactful 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。