Back to skill

Security audit

inquiry analyzer

Security checks for vulnerabilities and agentic risk

Overview

The skill is not clearly malicious, but it should be reviewed because it controls a logged-in browser, stores sensitive customer data locally, includes a hard-coded browser relay token, and can create undisclosed debug dumps.

Install only after reviewing the source and confirming the publisher is trusted for access to Alibaba inquiries and OKKI background data. Use a dedicated OpenClaw browser profile, do not run it on shared machines, avoid cloud-synced output folders, remove or rotate the hard-coded relay token, and clear chats/reports/debug-dumps after use. Ask for debug dumps to be opt-in and for sensitive outputs to use restricted permissions or redaction.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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 (3)

T09 · Insecure Skill Coding Practices

Warning
Location
lib/inquiry-analyzer.js:63
Finding

Hard-Coded OpenClaw Browser Relay Credential

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
lib/inquiry-analyzer.js:1439
Finding

Sensitive Customer Conversations and Enrichment Data Stored as Plaintext

Content
View full analysis
:"/\\|?*]/g, '_'); const chatFilename = `chats/${safeCustomerName}_${ts}.txt`; fs.writeFileSync(chatFilename, chatText); ``` It also writes generated reports without specifying restrictive permissions: ```js fs.writeFileSync(reportFile, md); ``` ```js fs.writeFileSync(csvFile, csv); ``` The OKKI workflow extracts additional customer information, including email and company data: ```js let email = ''; const emailMatch = ext.innerText.match(/[\w.+-]+@[\w.-]+\.\w+/); if (emailMatch) email = emailMatch[0]; let company = ''; const companyMatch = ext.innerText.match(/公司名称[::]\s*([^\n]+)/); if (companyMatch) company = companyMatch[1].trim(); return { name, country, email, company }; ``` The extracted information, background analysis, and enrichment details are subsequently written to Markdown and CSV report files: ```js fs.writeFileSync(summaryFile, summary); ``` ```js fs.writeFileSync(csvFile, csv); ``` ### Technical Analysis The Skill intentionally processes sensitive business and personal information from authenticated Alibaba and OKKI interfaces. This includes full conversations, customer names, countries, email addresses, company names, customer levels, inquiry identifiers, and background-check analysis. Persisting reports is part of the declared functionality. However, the implementation stores this data as ordinary plaintext and does not: - Set restrictive file permission ...[truncated 1794 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
lib/inquiry-analyzer.js:1095
Finding

Automatic Diagnostic Dumps Capture Excessive Authenticated Page Content

Content
View full analysis
{ const results = []; const scripts = document.querySelectorAll('script'); for (const script of scripts) { const content = script.textContent || ''; if (content.includes('productName') || content.includes('productTitle') || content.includes('subject')) { results.push(content.substring(0, 5000)); } } return results; }); if (jsonData.length > 0) { fs.writeFileSync(`${prefix}-json.json`, JSON.stringify(jsonData, null, 2)); console.log(` [Debug] JSON data saved: ${prefix}-json.json`); } } ``` ### Technical Analysis A routine data-extraction failure causes automatic collection of substantially more information than is required to produce the declared report. A full-page screenshot and co ...[truncated 2239 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (45)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill states it performs inquiry analysis, but it also launches child processes, resolves external script paths through environment variables, and can execute separate OKKI workflows. Indirect execution and path discovery increase the attack surface because behavior depends on external files and runtime environment that may not be visible in the top-level description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill states it performs inquiry analysis, but it also launches child processes, resolves external script paths through environment variables, and can execute separate OKKI workflows. Indirect execution and path discovery increase the attack surface because behavior depends on external files and runtime environment that may not be visible in the top-level description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill states it performs inquiry analysis, but it also launches child processes, resolves external script paths through environment variables, and can execute separate OKKI workflows. Indirect execution and path discovery increase the attack surface because behavior depends on external files and runtime environment that may not be visible in the top-level description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill states it performs inquiry analysis, but it also launches child processes, resolves external script paths through environment variables, and can execute separate OKKI workflows. Indirect execution and path discovery increase the attack surface because behavior depends on external files and runtime environment that may not be visible in the top-level description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill states it performs inquiry analysis, but it also launches child processes, resolves external script paths through environment variables, and can execute separate OKKI workflows. Indirect execution and path discovery increase the attack surface because behavior depends on external files and runtime environment that may not be visible in the top-level description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill states it performs inquiry analysis, but it also launches child processes, resolves external script paths through environment variables, and can execute separate OKKI workflows. Indirect execution and path discovery increase the attack surface because behavior depends on external files and runtime environment that may not be visible in the top-level description.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 308)May include surrounding context.

md
Relay Token 在主脚本中配置(`inquiry-analyzer.js`、`okki-background.js`),建议:

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 308)May include surrounding context.

md
Relay Token 在主脚本中配置(`inquiry-analyzer.js`、`okki-background.js`),建议:

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

A hard-coded relay token is embedded directly in source and used to connect to an authenticated browser relay. Secrets in code are easily leaked through source control, logs, backups, or redistribution, and in this context could let an attacker access or drive a logged-in browser session tied to Alibaba account data.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The code stores full chat transcripts and, in debug flows, saves screenshots, raw HTML, and embedded JSON from authenticated inquiry pages. Those artifacts can contain customer PII, business communications, account metadata, and page-resident tokens or identifiers, creating a much broader data exposure surface than the declared reporting purpose requires.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The debug function captures full-page screenshots, DOM HTML, and script-derived JSON from authenticated inquiry pages, all without warning or scoping. In this context, that can preserve sensitive customer and account data far beyond what is needed for troubleshooting, and the HTML/JSON may also include hidden identifiers and page state that increase downstream exposure.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The code writes chat histories containing customer data to local text files without disclosure, minimization, or access controls. Because the skill operates over authenticated sales inquiries, those transcripts may contain PII, commercial terms, and sensitive communications that become exposed to other local users, malware, backups, or later misuse.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The script’s declared purpose is inquiry analysis, but the implementation performs bulk OKKI background-check extraction and reporting on multiple inquiries up to a target item. This expands data access and processing beyond the stated scope, increasing the chance of unauthorized collection of personal and business intelligence without informed user approval.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This code actively clicks the "发起背调" control to trigger third-party background investigations, which is a substantive action rather than passive analysis. Initiating enrichment on contacts or companies can create privacy, compliance, and authorization risks if the user did not explicitly request or approve such processing.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The entrypoint and comments clearly describe an OKKI 'background investigation' workflow, which does not match the manifest's stated purpose of Alibaba inquiry analysis. This kind of capability mismatch is dangerous because it can conceal undisclosed data collection or execution behavior behind a benign-looking skill description, reducing user and reviewer visibility into what the skill actually does.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill documents access to environment variables and external path resolution but does not declare any tool scope, permissions, or allowed-tools boundary. This creates an authorization gap where the skill can rely on undeclared capabilities, making review and sandbox enforcement harder and increasing the chance of unintended access to host resources.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The file consistently presents the skill name, description, setup, usage, and warnings only in Chinese. This effectively forces a specific language on users without opt-in, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The documentation says the skill is self-contained and needs no external dependencies, but later requires an OpenClaw browser, gateway, login state, and external setup. Misrepresenting operational dependencies is dangerous because it obscures trust boundaries and the external systems and credentials the skill relies on.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Including OKKI background investigation in a skill framed as inquiry analysis expands the data-processing purpose into customer intelligence gathering. That creates privacy and least-privilege concerns because users may authorize the skill for reporting but not for collecting additional profiling information.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The documentation presents an inquiry-analysis skill, but these sections add OKKI background investigation and merged reporting capabilities beyond that narrow purpose. This is a scope-transparency problem that can lead to overcollection of sensitive customer or business intelligence under the guise of simple reporting.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill persists multiple local cache files and explicitly clears/rewrites them on each run, which goes beyond minimal report generation and introduces avoidable handling of local data artifacts. In a sales-inquiry analysis context, broad persistence increases privacy exposure and operational risk if the host machine is shared, compromised, or the files are later reused for unintended purposes.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The script automatically deletes files from several output directories on startup without confirmation. This is dangerous because it can cause unintended data loss, erase evidence or prior outputs, and in an automation environment may impact files the operator did not expect the skill to manage.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script writes complete chat history files using customer-derived filenames, preserving raw customer communications rather than only the structured fields needed for the report. In this skill's context, that constitutes unnecessary collection and persistence of customer data, increasing privacy, compliance, and insider-access risks.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The title filtering logic explicitly requires a product title to match an English-character regex and comments state it must be a 'pure English title', excluding Chinese, Spanish, French, German, Italian, Russian, and others. This is a natural-language locale policy constraint embedded in code, and there is no nearby opt-in, user choice, or documented justification for rejecting other languages.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The code parses and processes customer email plus company background data from the OKKI extension without any explicit disclosure or consent mechanism. In this skill context, that is more dangerous because the manifest frames the capability as inquiry analysis, not sensitive personal/business enrichment, so users may not expect this level of data handling.

Content

No source excerpt is available for this finding.