Back to skill

Security audit

雄韬企业头条助手

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent Chinese-language enterprise PR writing and poster-rendering helper, with disclosed local configuration storage and Playwright-based rendering but some dependency and privacy caveats.

Before installing, be comfortable with a Chinese-language corporate PR workflow that saves brand preferences locally, installs Playwright/Chromium from npm, and renders local HTML in a browser. For sensitive company messaging, avoid putting confidential material in the brand card unless local retention is acceptable, exclude `.company-headlines/` from source control, render only trusted HTML, and prefer pinning Playwright to a reviewed version.

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:10
Finding

Automatically Installed Dependency Is Not Version-Pinned

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:10-15 and SKILL.md:99-103
Vulnerability Type: Unpinned third-party dependency installation
Risk Level: Medium

Vulnerable Code

yaml
metadata:
  openclaw:
    requires:
      anyBins:
        - node
      install:
        - node: playwright

The dependency behavior is also documented as follows:

markdown
## Dependency Notes

When installing the Skill, OpenClaw automatically runs `npm install playwright` (including the Chromium browser).
Image generation invokes `node render.js` to render HTML as PNG.
No manual intervention, API key, or Python is required.

Technical Analysis

The Skill instructs OpenClaw to install playwright without specifying an exact reviewed version. No dependency lockfile or integrity hash is present in the audited project. Consequently, the code installed during a future deployment is not necessarily the same code that was implicitly trusted when this Skill was reviewed.

Package installation can execute third-party installation and lifecycle logic with the permissions of the OpenClaw process. Playwright also downloads or provisions substantial browser components. Although the audit found no evidence that the current package name is a typosquat or that Playwright itself is malicious, unconstrained dependency resolution creates a supply-chain exposure: a compromised registry account, malicious future release, registry substitution, or unexpected dependency-tree change could introduce executable code after review.

Attack Path

  1. An attacker compromises the package publication process, registry account, configured package registry, or a transitive dependency used by a future playwright release.
  2. The attacker publishes or serves a malicious package version under the dependency resolution accepted by npm install playwright.
  3. A user installs or reinstalls the Skill.
  4. OpenClaw automatica ...[truncated 873 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace the floating dependency with an exact, reviewed Playwright version rather than a package name that resolves to the latest release.
  2. Add and enforce a lockfile containing resolved versions and integrity hashes.
  3. Configure installation to use a trusted registry over authenticated TLS, and reject unexpected registry overrides.
  4. Use reproducible installation such as npm ci instead of unconstrained npm install.
  5. Disable dependency lifecycle scripts where operationally possible, or explicitly review every required lifecycle script before permitting it.
  6. Pin and verify downloaded browser binaries or use a pre-reviewed browser image with integrity validation.
  7. Run installation and rendering in a sandboxed, non-privileged account with restricted filesystem and network access.
  8. Add automated dependency scanning and require review before updating the pinned dependency or lockfile.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (29)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The declared description presents a high-level content-generation assistant for enterprise PR headlines and posters across multiple scenarios and industries. The supplied code does not implement any content generation, templating, industry presets, channel adaptation, or business logic for PR materials. Instead, it is narrowly a rendering utility that loads a provided local HTML file and captures it as a PNG via Playwright. While PNG poster export could be a supporting component of such a system, this code chunk alone materially differs from the declared purpose and omits the core advertised functionality.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The guide says the wizard writes .company-headlines/brand-card.md and that /头条 配置 will overwrite the existing file, but it does not mention confirmation, backup, or change warning. In an agent context, silent overwrite behavior can cause unintended loss of user-authored configuration or allow adversarial prompting to replace trusted brand settings without the user's informed consent.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The file is explicitly for a WeChat public-account format and all instructions are written to produce Chinese-language content, but it does not offer the user any language or locale choice. Under the policy rule, forcing a specific language without opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The instructions require drafting an internal announcement entirely in Chinese and specify Chinese enterprise messaging channels, but do not offer any language or locale choice. This is a natural-language policy concern because the skill implicitly enforces a specific language/locale rather than allowing user selection or documenting a justified locale constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This file is entirely written as a fixed Chinese prompt and instructs the model to produce a specific style of public announcement, with no indication that other languages are supported or that the user can opt in to Chinese output. Under the policy for natural-language violations, forcing a specific language without user choice is in scope.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The prompt instructs the model to write a WeChat public-account holiday greeting article entirely in Chinese and constrains the channel to '公众号版', without offering any user language or locale choice. This is a natural-language locale policy issue because the skill imposes a specific language/context by default rather than making it optional or explicitly justified as region-specific compliance.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The instructions require producing a WeChat public-account article in Chinese context and format, including platform-specific channel requirements, but do not offer any language or locale opt-in. Under the policy, forcing a specific language or locale without user choice is a natural-language policy violation unless clearly justified as region-specific, which is not stated here.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The renderer opens attacker-controlled or user-supplied HTML via Playwright and waits for network activity to settle, which allows the HTML to trigger outbound requests to external resources during rendering. In a screenshot-generation skill, this can leak IP/addressing metadata, perform unexpected network access, and create SSRF-like behavior if the HTML references internal services or sensitive local endpoints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The entire template is written as prescriptive instructions for producing channel-specific content in Chinese, including stylistic requirements and fixed wording conventions, but it does not offer any language or locale choice. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The entire template, headings, examples, and writing rules are written exclusively in Chinese and instruct the user to produce Chinese narrative content. There is no indication that the skill supports other languages, offers opt-in language selection, or is limited to a justified Chinese-only regional context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown template is entirely written in Chinese and presents a single-language output format without offering a language choice or explaining that it is intended only for a Chinese-language context. That creates a locale/language policy concern under the rule for natural-language policy violations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The entire template is written as a Chinese-only narrative format and provides no indication that language choice is optional or limited to a documented region-specific use case. This can violate language/locale policy because it implicitly requires 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
92% confidence
Finding

This markdown file presents all user-facing instructions and template guidance in Chinese, which can amount to a language/locale policy violation when no user opt-in or alternative language option is provided. The file does not state that the skill is region-specific or limited to Chinese-speaking users, so the forced locale appears undocumented.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

A language or locale policy violation applies when a skill forces a specific language without user opt-in. This README presents all usage, commands, and instructions only in Chinese, with no indication that users may choose another language or that the skill is intentionally limited to a Chinese-language context.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

For markdown files, this rule applies when the skill description omits warnings about behaviours that could affect user data, privacy, or system integrity. L65 instructs users to set GEMINI_API_KEY, but the README provides no caution about protecting the credential or that content generation depends on an external API service.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill states that the brand profile is automatically saved to .company-headlines/brand-card.md, but the lack of an explicit up-front warning can cause users to disclose sensitive company messaging, partner names, positioning, or internal preferences without realizing those details will persist locally. In enterprise contexts, silent persistence increases the chance of unintended retention, later exposure, or accidental inclusion in version control and backups.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

Automatically installing Playwright/Chromium introduces a substantial browser dependency and binary download surface that users may not expect from the description alone. While common, undisclosed installation of headless browser components can affect supply-chain risk posture, system footprint, and local execution trust decisions, especially in locked-down corporate environments.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The file presents all prompts, instructions, and invocation text only in Chinese, which can amount to a language policy constraint if users are not given an opt-in or alternative locale. The policy allows locale constraints when documented and justified, but this file does not state that the skill is intentionally limited to Chinese-speaking users or a China-specific context.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
98% confidence
Finding

Earlier in the document, the guide explicitly presents 7 questions (L003, L007-L119). However, the output section states '6 题完成后' before generating the brand card, which directly contradicts the documented flow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

该 markdown 文件适用 SQP-3。L146 明确将 LinkedIn 渠道限定为“英文/双语”,但文档中未说明这是可选项、需经用户确认,或基于特定地区/合规要求,因此构成潜在的语言/locale 策略约束。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

该 markdown 文件适用 SQP-3。L124-L130 明确要求用户通过中文命令“/头条 …”指定级别,整份技能说明也仅以中文交互为前提,未见用户可选择语言或明确限定为中文区域/组织场景的说明,存在默认强制特定语言的倾向。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

This markdown skill content appears to prescribe a single language/locale for the generated material, and there is no indication that users may opt into another language or that the restriction is required by a region-specific use case. Under the policy rules, forcing a specific language without user choice can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The entire prompt is written to produce a Chinese internal announcement and does not provide any opt-in or alternative language handling. Under the policy, forcing a specific language or locale without user choice can be a natural-language policy violation unless clearly justified as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The file’s title and all task instructions are written to produce a Chinese官网新闻稿, and there is no indication that users may choose another language or locale. Under the policy, forcing a specific language without opt-in can be a natural-language policy violation unless the regional constraint is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The top-level descriptive strings are written only in Chinese (HTML → PNG 渲染器, 用法), which imposes a specific language on users without any opt-in or indication that the tool is intended only for a Chinese-speaking audience. This can violate language/locale policy when no choice or justification is provided.

Content

No source excerpt is available for this finding.