ClawHub
Back to skill

Security audit

ClawGuard | Anti-Malware Scanner & Honeypot

Security checks for vulnerabilities and agentic risk

Overview

ClawGuard is a disclosed local monitoring skill whose elevated permissions fit its security-monitoring purpose, with shell actions gated on user approval.

Install only if you want a recurring local OpenClaw security monitor. Expect it to create a dummy honeypot file, scan your local OpenClaw skills, send alerts, and request approval before any kill or move quarantine action; read each alert before approving changes.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · skill.md (reported line 26)May include surrounding context.

md
## 2. Static Auditing (Read-Only)
- On your heartbeat schedule, use `local_read` to scan newly added directories in `~/.openclaw/skills/`.
- Flag any `SKILL.md` files that contain obfuscated `base64` strings, password-protected `.zip` extractions, or unrecognized `curl | bash` pipelines.

## 3. Incident Response & Approval Gate
- **MANDATORY POLICY:** You are strictly prohibited from autonomously executing shell commands. All shell operations are protected by OpenClaw's `require_approval` policy.

Session Persistence

Medium
Category
Rogue Agent
Confidence
84% confidence
Finding

This skill is configured to run on a recurring cron schedule and has both persistent local write access and shell capability, which creates a standing automated agent with the ability to change local state over time. Even though the text says shell use requires approval, the combination of scheduled execution, file creation, and broad permissions increases the blast radius if the skill is modified, subverted, or if approval controls fail.

Content

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

md
cron: "*/10 * * * *"
permissions: 
  - local_read
  - local_write
  - shell
  - notify
policies:

Static analysis

No suspicious patterns detected.