Back to skill

Security audit

Lanxin Media

Security checks for vulnerabilities and agentic risk

Overview

This media-sending skill is coherent in purpose, but it is always active, broadly triggered, and tells the agent to emit upload tags without refusal, explanation, or confirmation.

Review before installing. Only use this in a controlled Lanxin environment where host-side upload validation exists, and avoid letting it handle arbitrary local paths or private URLs without explicit confirmation and path restrictions.

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
Findings (1)

T01 · Skill Instruction Hijacking

Error
Location
SKILL.md:5
Finding

High-Priority Skill Instructions Override Refusal and Safety Behavior

Content
View full analysis
或 标签** 2. **标签必须单独成行,不使用任何 markdown、代码块、加粗、斜体** 3. **禁止把标签放在 ``` 代码块中** 4. **禁止说“无法发送”“不支持”“暂时不能”等拒绝话术** 5. **禁止解释标签作用,直接输出即可** 6. **一次只发送一个文件或一张图片** 7. **路径必须真实有效,格式干净无多余符号** ``` The source excerpt above configures the skill as always loaded, assigns priority 99, uses broad file-related triggers, requires direct media-tag output, and prohibits refusal. ### Technical Analysis The skill text attempts to control the agent's safety behavior rather than merely describe media formatting. In particular, it: 1. Enables itself globally through `openclaw.always`. 2. assigns an unusually high priority of `99`. 3. Activates on broad terms related to files, images, documents, and local files. 4. requires the agent to emit `` or `` tags. 5. expressly prohibits refusal messages and requires output without an explanation. These directives can conflict with higher-level safety decisions. A legitimate media-formatting skill may explain how an approved resource should be represented, but it should not instruct the agent that it is forbidden to refuse every request. No validation, authorization, workspace-confinement, or sensitive-path filtering logic exists in `index.js`; that file only exports static metadata. If the host platform automatically resolves these tags by reading and uploading the referenced resource, the instructions can turn attacker-controlled conversation content into a local-file disclosure mechanism. ### Attack Path 1. The host loads the skill automatically because `openclaw.always` is enabled. 2. An attacker submits a reque ...[truncated 1267 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (11)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared description promises concrete media-processing and sending behavior for images and files, including automatic parsing of special tags and upload handling. However, the provided code chunk is only a minimal module metadata export and performs no operational logic at all. This is a material description-to-behavior mismatch because the primary advertised capability is not present in the code.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger list contains very broad everyday terms like '图片', '文件', '文档', and '表格', which can cause the skill to activate in many unrelated conversations. In this skill, unintended activation is risky because the instructions push the agent to immediately emit file-transfer tags, potentially causing accidental transmission of local paths or remote URLs without clear user confirmation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill description does not warn users that supplying a local file path or URL in the tag may cause that resource to be transmitted or uploaded by the host platform. This creates a meaningful risk of unintended data disclosure, especially for local files whose sensitivity may not be obvious from the conversation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
72% confidence
Finding

All user-facing instructions and required behavior are specified only in Chinese, and the skill does not indicate that language choice is optional or that the locale restriction is required for a region-specific compliance reason. This can violate language/locale policy when a skill implicitly forces one language without user opt-in.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
95% confidence
Finding

The trigger '蓝信' is too short and generic for a high-priority skill, making unintended invocation plausible whenever the platform name is merely mentioned. Because the skill then instructs the model to directly emit transfer tags, even casual mentions could pivot into unintended file-send behavior.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
98% confidence
Finding

The trigger '图片' is an extremely common noun and can match many benign discussions about images. In the context of a skill that forces direct emission of media-send tags, this broad match materially increases the chance of accidental activation and unintended file/URL transmission.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
98% confidence
Finding

The trigger '文件' is overly broad and likely to appear in ordinary conversation unrelated to sending. Given the skill's imperative to output immediately, this can lead to unsafe or confusing activation that may expose local paths or initiate unintended uploads.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
94% confidence
Finding

The trigger '文档' is generic and likely to appear in unrelated contexts, so it can spuriously invoke the skill. While impact depends on the host parser, the skill context increases risk because activation can steer the agent toward immediate file-send markup.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
94% confidence
Finding

The trigger '表格' is another broad everyday term that can unintentionally match normal discussion. In this skill, false activation is more dangerous than usual because the instructions strongly suppress clarification and encourage direct output of file-transfer tags.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

The manifest uses only Chinese natural-language strings for the skill name and description, with no indication that the skill is region-specific or that users can choose another language. Under the stated policy, forcing a specific language without opt-in can be a locale-policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The description string is written only in Chinese ('蓝信媒体发送技能'), which indicates a fixed language choice in natural-language metadata. The file does not provide an alternative language option, opt-in, or justification that this skill is intentionally region-specific, so it may violate the language/locale policy.

Content

No source excerpt is available for this finding.