Back to skill

Security audit

Openclaw Chat

Security checks for vulnerabilities and agentic risk

Overview

The package looks like an OpenClaw mobile admin-panel prototype, but it overclaims high-impact remote management and repair features that are not safely scoped or actually implemented.

Treat this as a prototype or mock admin UI unless the publisher supplies a real, authenticated backend and clear controls for memory sync, remote management, destructive actions, and dependency updates. Do not rely on its repair, stop/start, restore, or cron screens for real operations as currently packaged.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (62)

Known Vulnerable Dependency: @babel/plugin-transform-modules-systemjs==7.29.0 — 1 advisory(ies): CVE-2026-44728 (@babel/plugin-transform-modules-systemjs generates arbitrary code when compiling)

High
Category
Supply Chain
Confidence
88% confidence
Finding

The project includes @babel/plugin-transform-modules-systemjs 7.29.0, which is reported to generate arbitrary code during compilation. In a frontend build chain, this is especially risky if the toolchain compiles untrusted or externally supplied code, because compromise would occur before deployment and could taint shipped artifacts.

Content

No source excerpt is available for this finding.

Known Vulnerable Dependency: @xmldom/xmldom==0.8.12 — 14 advisory(ies): CVE-2026-83608 (xmldom: DocType `name` Injection Bypasses requireWellFormed); CVE-2026-41673 (xmldom: Uncontrolled recursion in XML serialization leads to DoS); CVE-2026-83605 (xmldom: Attribute name injection via setAttribute() bypasses requireWellFormed) +11 more

High
Category
Supply Chain
Confidence
86% confidence
Finding

@xmldom/xmldom 0.8.12 is flagged with multiple XML parsing/serialization advisories, including injection and denial-of-service issues. Because this dependency is used by plist handling in the Capacitor/native tooling chain, malicious XML/plist input could affect local tooling, CI jobs, or packaging workflows.

Content

No source excerpt is available for this finding.

Known Vulnerable Dependency: brace-expansion==5.0.5 — 4 advisory(ies): CVE-2026-13149 (brace-expansion: DoS via exponential-time expansion of consecutive non-expanding); CVE-2026-45149 (brace-expansion: Large numeric range defeats documented `max` DoS protection); CVE-2026-14257 (brace-expansion: DoS via unbounded expansion length causing an out-of-memory pro) +1 more

High
Category
Supply Chain
Confidence
90% confidence
Finding

brace-expansion 5.0.5 is associated with multiple denial-of-service issues involving pathological expansion patterns and memory growth. In build or CLI tooling that evaluates attacker-controlled glob-like input, this can crash or significantly delay CI and developer workflows.

Content

No source excerpt is available for this finding.

Known Vulnerable Dependency: browserslist==4.28.2 — 2 advisory(ies): CVE-2026-73088 (Browserslist: Uncaught crash / prototype write via untrusted browserslist-stats.); CVE-2026-73089 (Browserslist: Unbounded memory growth (no cache eviction) via distinct query res)

High
Category
Supply Chain
Confidence
84% confidence
Finding

browserslist 4.28.2 is reported with crash/prototype-write and unbounded-memory issues when fed untrusted stats or distinct queries. Since this project uses modern frontend tooling that depends on browserslist during builds, the primary risk is to build integrity and availability in local or CI environments rather than direct production runtime exploitation.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
81% confidence
Finding

The description indicates automatic remote/local memory sync and persistent multi-agent collaboration, implying session persistence and data retention across contexts. Without any disclosure of retention scope, sync boundaries, access controls, or user consent, sensitive prompts, agent memory, or operational metadata could persist longer than expected or be exposed across devices/services.

Content

Scanner excerpt · SKILL.md (reported line 8)May include surrounding context.

md
## Features

1. **OpenClaw Native Chat UI** - Mobile interface designed specifically for OpenClaw
2. **Agent Team Group Discussion** - Create groups for multi-agent collaboration
3. **Remote & Local Memory Sync** - Agent memory syncs automatically
4. **Integrated Remote Management UI** - Manage local services from anywhere
5. **Local Port & App Status Management** - Monitor and control local service status in real-time

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The skill advertises remote service management, real-time status control, cron management, and automatic memory synchronization without any warning about the security and operational consequences. In this context, users could enable capabilities that alter local services, scheduled tasks, or move data between local and remote contexts without understanding trust boundaries, authentication needs, or failure impact.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The UI exposes high-risk operational actions such as starting/stopping OpenClaw, running doctor --fix, restoring a stable config, restarting the gateway, and triggering repair directly from click handlers without any confirmation dialog, re-authentication, or explicit risk acknowledgement. In an admin/control-panel context, accidental taps, clickjacking, or misuse by anyone with UI access could cause service disruption or unintended recovery actions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The UI presents controls such as 'Start OpenClaw', 'Stop OpenClaw', 'Restart Gateway', and 'doctor --fix' as if they perform real operational actions, but the handlers only mutate local React state and show success toasts via setTimeout. This can mislead operators into believing critical recovery or service-management actions succeeded when nothing actually happened, creating dangerous false assurance during incident response or outage handling.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The interface exposes destructive or high-risk actions such as stopping services, restoring stable config, deleting services, and triggering repair without any confirmation dialog, acknowledgment, or rollback warning. In a real deployment, a stray click or confused operator could interrupt service, overwrite configuration, or remove entries immediately.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

This code presents buttons for potentially destructive operations such as stopping OpenClaw, running doctor --fix, restoring a stable config, restarting the gateway, and deleting services, but the handlers execute immediately on click with no confirmation prompt or explicit warning message. For a code file, these are safety-relevant operations that can affect system availability or configuration and should include some user disclosure before execution.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 1642)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 3315)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 3317)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 5385)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 5389)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 5611)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · package-lock.json (reported line 5613)May include surrounding context.

json
"kleur": "^4.1.5",
        "native-run": "^2.0.3",
        "open": "^8.4.0",
        "plist": "^3.1.0",
        "prompts": "^2.4.2",
        "rimraf": "^6.0.1",
        "semver": "^7.6.3",

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · android/app/src/main/assets/public/assets/index-DqHmsSsP.js (reported line 175)May include surrounding context.

js
agentList: 'Agent 列表',
    scheduledTasks: '定时任务',
    recentLogs: '最近日志',
    groupList: '群组列表',
    searchPlaceholder: '搜索...',
    nextRun: '下次',
    port: '端口',

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · android/app/src/main/assets/public/assets/index-DqHmsSsP.js (reported line 179)May include surrounding context.

js
agentList: 'Agent 列表',
    scheduledTasks: '定时任务',
    recentLogs: '最近日志',
    groupList: '群组列表',
    searchPlaceholder: '搜索...',
    nextRun: '下次',
    port: '端口',

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · dist/assets/index-aX36dzTn.js (reported line 175)May include surrounding context.

js
agentList: 'Agent 列表',
    scheduledTasks: '定时任务',
    recentLogs: '最近日志',
    groupList: '群组列表',
    searchPlaceholder: '搜索...',
    nextRun: '下次',
    port: '端口',

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · dist/assets/index-aX36dzTn.js (reported line 179)May include surrounding context.

js
agentList: 'Agent 列表',
    scheduledTasks: '定时任务',
    recentLogs: '最近日志',
    groupList: '群组列表',
    searchPlaceholder: '搜索...',
    nextRun: '下次',
    port: '端口',