ClawHub
Back to skill

Security audit

AgentMemory Adapter

Security checks for vulnerabilities and agentic risk

Overview

This skill openly helps install and harden AgentMemory, including persistent memory and agent configuration changes, with no evidence of hidden or malicious behavior.

Install this only if you want AgentMemory to persist coding-agent context and modify Codex/OpenClaw configuration. Use localhost or HTTPS for any bearer-token-protected server, avoid storing secrets in memory, enable isolated agent scope for sensitive or multi-role work, and review the generated config and hooks after setup.

SkillSpector

Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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)

Lp3

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding
The skill clearly instructs the agent to read local reference files, inspect and modify agent configuration, access environment variables such as AGENTMEMORY_SECRET, and communicate with a local or remote AgentMemory service, but it does not declare any permissions for those capabilities. This creates a permission-transparency gap: a user or platform may treat the skill as lower risk than it really is, increasing the chance of unintended file access, network calls, or secret exposure during execution.

Session Persistence

Medium
Category
Rogue Agent
Content
---
name: agentmemory-adapter
description: Install, wire, audit, and harden AgentMemory for AI coding agents and project workspaces. Use when Codex needs to adopt rohitg00/agentmemory, configure Codex or OpenClaw memory integration, create ClawHub-ready AgentMemory guidance, debug MCP/hook/plugin setup, verify memory capture, or handle edge cases such as npx stalls, sandbox localhost routing, duplicate plugin paths, stale Codex hooks, remote bearer-token security, multi-agent isolation, or missing full-tool proxy mode.
license: MIT
metadata:
  openclaw:
Confidence
76% confidence
Finding
This skill is specifically designed to install and operationalize a shared memory layer across coding agents and workspaces, including verifying memory capture and configuring shared or isolated scopes. Persistent memory can retain prompts, code, secrets, or project context beyond the current session, which creates privacy and cross-context leakage risk if retention, isolation, and deletion controls are not explicit and correctly configured.

Static analysis

No suspicious patterns detected.