Back to skill

Security audit

OpenClaw 全岗位技能解释器

Security checks for vulnerabilities and agentic risk

Overview

This skill is mainly a Chinese glossary, but it tells the agent to automatically install and update matched skills without clear scope or risk review.

Review this before installing. It can help Chinese-speaking users understand skill names, but do not rely on its safety claims. Require an explicit review of the exact matched skill, publisher, permissions, data access, and rollback path before allowing any install or automatic update, especially for scraping, proxy, email, calendar, meeting, or note-related skills.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (10)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The title and description state the skill is for domestic users who do not understand English and that it translates skill names into plain Chinese. This imposes a specific language/locale behavior without mentioning any option for users who may prefer another language.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill states that after the user confirms, it will automatically install a matched skill, but it does not warn that installation changes the system's available capabilities and may introduce sensitive behaviors. Because this skill is positioned as a plain-language translator for non-technical users and includes security-related skills, users may approve installation without understanding the permissions, trust implications, or operational impact.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The invocation examples are extremely broad (e.g., asking for tools that automate PPT creation or scrape web data), which makes it unclear when this skill should activate and what downstream skill it may select. In a skill that can lead to automatic installation, ambiguous matching increases the chance of unintended activation and installation of capabilities the user did not specifically request.

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 glossary and its instructional framing are written only in Chinese, including statements such as '说大白话就能精准匹配对应技能', with no indication that users may choose another language or that the skill is limited to a Chinese-speaking context. Under the policy, forcing a specific language without opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document repeatedly claims that it contains only 'verified safe skills' and 'no risk/no backdoor', while also listing capabilities with clear abuse potential such as proxy management, automated updating, broad tool orchestration, and privileged execution concepts. This creates a false sense of safety that can lower reviewer scrutiny and encourage deployment of risky skills without proper threat modeling or controls.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The entries for meeting recording, email automation, calendar sync, and note sync describe handling potentially sensitive communications and personal data without warning about consent, retention, access scope, or third-party transfer. Users may enable these functions without understanding the privacy implications, leading to over-collection, unauthorized processing, or compliance violations.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

Describing a 'network proxy manager' as an officially certified safe skill while stating that it helps scraping avoid being blocked normalizes evasion behavior and can facilitate bypassing platform restrictions or monitoring. In this context, the safety claim increases danger because it masks a capability commonly associated with abuse, policy violations, or unauthorized collection.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The manifest description states the skill is a Chinese glossary/explainer and matches skills using colloquial Chinese, which implies a fixed language/locale behavior. There is no accompanying indication that users can choose another language or that the Chinese-only scope is a documented, region-specific requirement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The comment '自然语言匹配技能' and the surrounding logic using cnName suggest the skill is designed around Chinese-language matching behavior. There is no visible indication in this file that users can choose language/locale or that the Chinese-specific behavior is explicitly opt-in, which may conflict with language-choice policy expectations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The trend-analysis description mentions automated crawling of industry and competitor data without noting legal, contractual, privacy, or robots/terms-of-service constraints. While the text is high level, presenting crawling as frictionless can encourage noncompliant collection practices or unnoticed external interactions.

Content

No source excerpt is available for this finding.