Security audit

免费版

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent data-analysis helper, but it asks users to upload potentially sensitive datasets while giving misleading privacy assurances and includes unrelated Telegram promotion in active skill instructions.

Install only if you are comfortable with uploaded dataset contents being processed by your configured AI provider. Do not rely on the package’s “not uploaded to third parties” wording unless your OpenClaw setup is actually local-only. Expect broad natural-language activation, and be aware that the active skill text includes unrelated Telegram promotion.

A.I.G

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 (2)

T01 · Skill Instruction Hijacking

Error

Location: SKILL.md:162
Finding: Unrelated Telegram Promotion Embedded in Active Skill Instructions
Content: View full analysis
Remediation: View remediation

other

Note

Location: SKILL.md:54
Finding: Misleading Claims About Third-Party Data Processing
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (7)

Vague Triggers

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger list includes very generic phrases such as 'find insights', 'data report', '看看这个数据', and '帮我分析' plus auto-triggering on common file uploads. This can cause the skill to activate in unrelated conversations and process uploaded business data without sufficiently explicit user intent, increasing the risk of accidental data exposure to the configured AI backend.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The skill encourages users to upload sensitive files for analysis but does not clearly and prominently warn that file contents may be sent to the user's configured AI provider. Although the FAQ mentions data flows between OpenClaw and the configured AI, the top-level description does not provide a conspicuous consent/privacy warning appropriate for financial, sales, and user-behavior datasets.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The line states "中文数据 → 全程中文输出," which imposes a language choice based on detected data characteristics rather than user preference. This is reinforced later by a rule requiring the entire report in Chinese for Chinese data, with no opt-in or alternative language option.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: Rule 4 says that for Chinese data, the entire report must be output in Chinese. This is a language-policy issue because it hard-codes locale behavior based on dataset content and does not offer the user a choice of output language.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The marketing copy encourages users to upload Excel files and broadly claims the data 'will not be uploaded to any third party' and is 'safe' without disclosing actual data-handling boundaries, model-provider exposure, logging, or deployment-dependent risks. This can mislead users into sharing sensitive business data under false assumptions, especially because the skill relies on external AI providers and may process confidential spreadsheet contents.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The example trigger phrase “帮我分析这个数据，找出销售下滑原因” is presented as the way to use the skill, but the description does not define any bounded trigger conditions or exclusions. Phrases like “帮我分析这个数据” are broad everyday requests that could match many unrelated contexts, increasing the risk of unintended invocation in systems that rely on natural-language routing.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The trigger phrases are very broad, generic natural-language requests such as '帮我分析这个数据' and '这个 Excel 有没有异常数据', which are likely to overlap with ordinary user conversation. In an agent/skill environment, this can cause unintended invocation on sensitive datasets or in contexts where the user did not explicitly mean to activate this skill, increasing the risk of accidental data processing or disclosure to the configured model provider.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.