ClawXMemory

Security checks across malware telemetry and agentic risk

Overview

ClawXMemory is a coherent memory plugin, but it stores full conversations and appears to use environment-based credentials for network LLM extraction without clearly declaring that credential use.

Install only if you want OpenClaw conversations stored as persistent memory. Before enabling it, check which LLM provider credentials are available in your environment, keep the dashboard on 127.0.0.1, and adjust or disable full-session capture, auto-indexing, add, and recall if that is broader than you want.

VirusTotal

VirusTotal engine telemetry is currently stale for this artifact.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

#
ASI03: Identity and Privilege Abuse
High
What this means

API keys or other environment secrets may be used to send memory-extraction requests to an external provider, and that credential pathway is not clearly visible in the declared install contract.

Why it was flagged

This shows the LLM extraction path can read an environment-named secret and use it in network activity, while the registry requirements list no required env vars, env var declarations, or primary credential.

Skill content
Static scan at line 2013: "Environment variable access combined with network send"; code snippet: "process.env[configured]"
Recommendation

Declare the supported credential/env-var names and provider destinations, require explicit user configuration, and run the plugin only with narrowly scoped provider credentials.

#
ASI06: Memory and Context Poisoning
Medium
What this means

Private conversation details may be stored in the local SQLite memory and later recalled into future agent context.

Why it was flagged

The plugin persists memory locally and defaults to capturing full sessions, including assistant messages; this is central to a memory plugin but means sensitive conversation context can be retained and reused.

Skill content
"dataDir": { "description": "Base directory used to persist local memory data." }, "captureStrategy": ... "default": "full_session", "includeAssistant": { "default": true }
Recommendation

Review the configured data directory, use clear/export controls carefully, and disable or narrow capture/recall settings if you do not want broad persistent memory.

#
ASI10: Rogue Agents
Low
What this means

The plugin may continue updating and using memory without a manual indexing command each time.

Why it was flagged

Periodic indexing and automatic add/recall are enabled by default. This is disclosed and purpose-aligned, but it is background persistence that users should understand.

Skill content
"autoIndexIntervalMinutes": { "type": "integer", "default": 60 }, "recallEnabled": { "default": true }, "addEnabled": { "default": true }
Recommendation

Set the auto-index interval to 0 or disable add/recall if you only want manual memory use.

#
ASI07: Insecure Inter-Agent Communication
Low
What this means

Anyone with access to the local machine/session may be able to view the dashboard and inspect stored memory.

Why it was flagged

The plugin starts a local dashboard by default. It is bounded to loopback and described as read-only, but it can display stored memory data to local browser/process access.

Skill content
"uiEnabled": { "default": true }, "uiHost": { "default": "127.0.0.1" }, "uiPort": { "default": 39393 }, "help": "Start local read-only dashboard server"
Recommendation

Keep the UI host bound to 127.0.0.1 and disable the dashboard if you do not need it.