ClawHub
Back to skill

Security audit

nexus-memory

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent persistent-memory skill, but it automatically stores and reuses conversation-derived facts in a shared memory database with limited scoping or consent controls described.

Install only if you are comfortable with conversation-derived facts being stored in Qdrant and reused across sessions and compatible agents. Avoid using it around secrets, credentials, regulated data, or sensitive client material unless you have clear controls for review, deletion, namespacing, and disabling automatic capture.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill states that it automatically searches memories before each prompt and stores notable facts after each turn, but it does not warn users that transcript-derived data will be persisted and later re-injected into future conversations. This creates a significant privacy and data-handling risk because users may disclose secrets or sensitive project details without realizing they will be retained and surfaced later.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The activation guidance is broad enough that the skill may be invoked whenever past context might be useful, causing automatic recall or capture in situations the user did not clearly intend. Because this skill injects memories before prompts and stores facts after turns, over-broad triggering increases the chance of unnecessary data exposure and persistence.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill advertises a shared Qdrant collection across Claude Code, Hermes, OpenClaw, and any MCP-compatible agent without clearly warning that memories written by one agent become visible to others. This cross-agent visibility can lead to unintended disclosure of sensitive information across tools, trust boundaries, or user contexts.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.