Back to skill

Security audit

Subagent Setup Wizard

Security checks for vulnerabilities and agentic risk

Overview

This Subagent setup guide is not malware, but it may surface broad operational instructions that change or disrupt an OpenClaw environment without enough guardrails.

Install only if you want a Chinese-language Subagent configuration guide and are comfortable reviewing commands before use. Treat any config set, config reset, gateway restart, and subagents kill commands as manual administrative actions that may affect active OpenClaw sessions or resource use.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger conditions are broad enough to activate on generic mentions of subagent, 并发, or 多 agent, which could cause this skill to be selected in unrelated contexts. Because the skill contains operational commands and configuration guidance, over-triggering increases the chance that risky instructions are surfaced when not appropriate to the user's actual intent.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill includes executable shell scripts and direct configuration-changing commands (config set, config reset, restart, kill) even though it is framed as a setup/advisory wizard. That makes the skill capable of driving operational changes on a host or agent environment, increasing the risk of accidental destructive actions or abuse if the content is invoked automatically or copied without review.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The troubleshooting section recommends disruptive commands like configuration reset, gateway restart, and killing subagents without warning about service interruption, loss of state, or impact on active work. Users may execute these commands verbatim during troubleshooting, causing avoidable downtime or data/session loss.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The description is written only in Chinese ("Subagent 启用配置向导"), which can impose a language-specific experience without indicating user choice or opt-in. Under the policy, locale or language constraints should be optional or clearly justified when they limit accessibility.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

In the '研究分析任务' example, the skill recommends creating subagents to '抓取官方文档' and analyze GitHub source code. External documentation retrieval and repository analysis go beyond the manifest's stated focus on version introduction, enabling configuration, concurrency best practices, and resource monitoring.

Content

No source excerpt is available for this finding.