registry / @xmem.space/openclaw-plugin / 8.1.1

@xmem.space/openclaw-plugin@8.1.1

OpenClaw collector plugin for xmem — graph-based long-term memory for AI agents. Deep workspace capture (all relevant file types, 4 levels deep, periodic re-scan every 2h). Multi-agent aware with per-agent Space/Key routing. Triple-ingest (artifacts + att

AI Security Review

scanned 2h ago · by lpm-firewall-ai

LPM flags this version as an AI-agent control-surface risk. Installing the npm package modifies the user-wide OpenClaw allowlists before the user configures or invokes the extension. Once enabled and configured, the extension can upload scanned workspace content to its service.

Static reason

One or more suspicious static signals were detected.

Trigger

npm installation runs postinstall; later OpenClaw bootstrap or plugin activity can activate workspace capture.

Impact

The package grants its own agent tool and plugin access through a foreign global configuration file without consent, enabling an extension that can collect workspace data.

Mechanism

Postinstall allowlist mutation and workspace-content upload.

Policy narrative

The package’s npm postinstall script silently edits the user’s OpenClaw configuration, adding memtap to both tool and plugin allowlists. This grants the extension access without an explicit OpenClaw configuration action. Its runtime code can then scan workspace files and send their paths and content to the configured xmem API.

Rationale

This is unconsented postinstall mutation of a user-wide AI-agent control surface. The later workspace upload capability makes that automatic enablement security-significant.

Evidence

package.jsonscripts/post-install.shindex.ts${HOME}/.openclaw/openclaw.json.xmem-hashes.json

Network endpoints1

api.xmem.space

@xmem.space/openclaw-plugin@8.1.1

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code