ClawHub
Back to skill

Security audit

Multi Agent Blueprint

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed setup guide for multi-agent OpenClaw deployments, but users should handle credentials and persistent memory carefully.

Install only if you want a multi-agent OpenClaw blueprint. When applying it, use least-privilege bot/API/SSH credentials, keep secrets out of agent files and memory, restrict Telegram access with allowlists and mentions, and define retention/deletion rules for persistent memory.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Session Persistence

Medium
Category
Rogue Agent
Content
*Social agents can use specialized providers like xAI/Grok for platform-native tone.

### 2. Create Agent Directories

```bash
# For each agent:
Confidence
60% 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.

Session Persistence

Medium
Category
Rogue Agent
Content
}
```

### 4. Create Agent Files

Each agent needs 4 files in their `agentDir`:
Confidence
80% 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.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The document recommends storing API keys in .env files and centralizing SSH/NAS credentials in a single agent, but it omits basic safeguards such as least privilege, file permission hardening, rotation, segregation by agent, and avoiding exposure through logs or memory. Because this blueprint is for multi-agent deployments with cross-agent routing and external integrations, weak credential handling could lead to broader compromise if one workspace, agent, or host is exposed.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill explicitly instructs agents to persist 'key decisions, dates, and action items' into memory files before session compaction, but it does not warn that these files may contain sensitive or regulated data and will survive session resets. In a multi-agent environment handling finance, health, contracts, and personal data, this creates a realistic risk of over-retention, unintended disclosure, and sensitive context being resurfaced later via memory search.