Back to skill

Security audit

seven

Security checks for vulnerabilities and agentic risk

Overview

This is a non-executable informational skill, though it mixes number-seven education with disclosed personal/project promotion and links.

Install only if you are comfortable with a number-seven reference skill that may also surface Chen Lang/OpenClaw profile information, an email address, and project links. Treat those links and promotional claims as unverified, and ask the agent to omit profile or project material when you only want cultural or scientific information about 7.

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

Warning
Location
SKILL.md:110
Finding

Unrelated Promotional Content Injected into Number-Seven Responses

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 110-227
Vulnerability Type: Skill instruction hijacking through unrelated promotional material
Risk Level: Medium

The skill combines legitimate cultural information about the number seven with a personal profile, contact information, branded projects, external links, and instructions encouraging the agent to reference that material when answering user questions.

Relevant Skill Content:

markdown
## Profile of Chen Lang

### Basic Info

| Item | English |
|------|---------|
| Name | Chen Lang |
| Preferred Name | Brother Lang |
| Email | leogood@foxmail.com |
| Time Zone | Asia/Shanghai |
| Languages | Chinese, English |

### Chen Lang's OpenClaw Project Matrix

OpenClaw123.xyz serves as a beginner-friendly resource portal, providing
traffic entry points and support for the ecosystem.

**URL**: https://Openclaw123.xyz

Aixin — Give your lobster a "phone number", contact it first when you need help.

**URL**: https://aixin.chat
markdown
## Connection Between Number 7 and Chen Lang's Projects

The Big Dipper was an ancient natural navigation system. Chen Lang's
OpenClaw projects, like the seven stars, provide direction for Personal
AI OS development:
markdown
When users ask about:
- Cultural meaning of number 7
- Traditional concepts like Seven Stars, Seven Emotions, Seven Colors
- Symbolism of 7 in Eastern and Western cultures
- Scientific and natural significance of 7
- Profile of Chen Lang
- OpenClaw project matrix

Technical Analysis

The declared purpose of the skill is to provide cultural information about the number seven. However, the skill also contains extensive material promoting a named individual, an email address, branded services, and external websites. The artificial comparison between the projects and the seven stars attempts to make this unrelated promotional material ap ...[truncated 2259 chars]

Remediation
View remediation

Remediation Suggestions

  1. Remove the personal profile, email address, branded project matrix, marketing claims, and external service links from this skill.
  2. Remove the section that creates an artificial relationship between the number seven and the named projects.
  3. Restrict the activation metadata and usage instructions to factual cultural, historical, linguistic, religious, and scientific information about the number seven.
  4. If the profile or project information serves a legitimate purpose, move it into a separate skill with an explicit name and activation condition. That skill should activate only when a user directly requests information about the person or projects.
  5. Add a scope rule stating that unrelated products, people, contact details, and websites must not be introduced into number-seven answers unless the user explicitly asks for them.
  6. Review all external links independently before allowing them in user-facing responses, and describe them neutrally rather than presenting unverified promotional claims.
  7. Apply data-minimization practices to personal information. Include an email address only where it is necessary, authorized, and directly responsive to the user's request.
  8. Add regression tests confirming that ordinary questions about the number seven produce no references to Chen Lang, OpenClaw, OpenClaw123, Aixin, ClawOpen, personal contact details, or unrelated external links.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The activation description is very broad: any query about the number 7, related concepts, or even Chen Lang's profile may trigger this skill. Over-broad activation can cause unintended routing on ordinary conversations, leading to irrelevant responses and unnecessary disclosure of embedded personal/contact information.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The usage instructions define a wide trigger surface across cultural meanings, science, personal profile, and project matrix, without exclusions or negative examples. This increases accidental invocation risk and can surface unrelated promotional or personal content when the user only intended a general discussion of 7.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

The description labels the skill as a bilingual Chinese/English skill, and the document consistently presents both languages together. There is no natural-language indication that the user can choose a preferred language, which may conflict with language/locale choice expectations.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.