ClawHub
Back to skill

Security audit

openclaw-behavior-plan

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only planning skill that may over-trigger or produce Chinese-only plans, but it does not install code, persist itself, or perform actions on its own.

Install this only if you want an OpenClaw planning helper and are comfortable with Chinese-language output. Review generated plans before executing them, especially when they include shell commands, file writes, Slack/email/calendar actions, or other external side effects.

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
96% confidence
Finding
The skill description is broadly scoped to common planning requests like creating plans, designing behavior, and multi-step task execution, which increases the chance of over-triggering on ordinary user requests. Overly broad activation can cause the wrong skill to take control, leading to unintended planning behavior, unnecessary tool-oriented decomposition, or bypass of more specific and safer skills.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
The skill content is written to force Chinese output and includes a fixed Chinese trigger phrase, without offering language negotiation or documenting a strict locale requirement. This can cause user intent mismatch, reduce operator visibility during review, and increase the risk that important safety constraints or sensitive-step confirmations are misunderstood or skipped in multilingual environments.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The example workflow performs an external side effect by posting to Slack based on derived calendar data, but it does not include any explicit user confirmation step immediately before sending the message. In an agent-planning skill, examples strongly shape generated plans, so this omission can normalize autonomous outbound actions and cause the agent to contact third parties with incorrect, sensitive, or unintended meeting proposals.

Static analysis

No suspicious patterns detected.