ClawHub
Back to skill

Security audit

记忆编排器

Security checks across malware telemetry and agentic risk

Overview

This memory-management skill is mostly purpose-aligned, but it asks for command execution and persistent memory handling without enough concrete scoping or user-control safeguards.

Review this before installing if you expect strict control over local command execution or retained personal/work memory. Use it only in an environment where persistent memory files are acceptable, avoid storing secrets, and require explicit confirmation for cleanup, deletion, archival, and any shell command execution.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (3)

Intent-Code Divergence

Medium
Confidence
90% confidence
Finding
The skill claims command execution is restricted to a whitelist, but the file exposes the exec tool and provides no concrete command validation, allowlist definition, or enforcement logic. In an agent ecosystem, unsupported safety claims can cause operators or downstream agents to trust and invoke exec in unsafe ways, increasing the chance of arbitrary command execution or unsafe local actions.

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill advertises read, write, and exec capabilities for a memory-management function, which is broader than necessary and creates a larger attack surface. In agent environments, unnecessary exec access can be abused through prompt-driven workflows, persistence operations, or future extensions to run local commands against the host system.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill supports persistent storage, auto-cleanup, archival, and possible deletion of memory data without a clear, prominent warning or explicit consent model for retention and deletion behavior. For agent memory, this can lead to silent storage of sensitive user data or unintended loss/archival of important context, especially because 'important memory' may include identity and preference information.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.