ClawHub
Back to skill

Security audit

精英长记忆

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed long-term memory skill, but it should be reviewed carefully because it automatically stores user context across sessions without clear consent, retention, or deletion controls.

Install only if you want the agent to keep durable project memory in workspace files. Before use, decide what may be stored, avoid entering secrets or regulated personal data, and periodically inspect or delete SESSION-STATE.md, MEMORY.md, and memory/ logs. Treat callback_url as unsafe to use unless you know exactly what data will be sent and where.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (4)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to automatically read and persist conversation context across sessions in local files without an explicit user-facing warning or consent model. In this context, the stored data includes preferences, decisions, deadlines, and corrections, creating a realistic privacy and data retention risk if sensitive information is logged by default or later exposed to other users/processes with workspace access.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The documented callback_url parameter implies that outputs or status may be sent to an external endpoint, but the skill provides no warning about data leaving the local environment. Without explicit disclosure, users may unknowingly route sensitive task content or metadata to third-party services, creating exfiltration and compliance risk.

Ssd 3

Medium
Confidence
96% confidence
Finding
The skill explicitly promotes cross-session retention of natural-language user context in SESSION-STATE.md, MEMORY.md, and dated logs. In an agent skill with read/write capabilities, this is dangerous because it normalizes broad retention of potentially sensitive conversation content, increasing the chance of privacy leakage, unintended reuse, or exposure through the workspace, backups, or source control.

Ssd 3

Medium
Confidence
97% confidence
Finding
The 'write before reply' workflow creates a default logging requirement for many categories of user input, including preferences, decisions, deadlines, and corrections. In this skill’s context, that makes the privacy risk more acute because persistence is mandatory in normal operation rather than exceptional, so sensitive or regulated data can be captured reflexively before any review or consent check occurs.

Static analysis

No suspicious patterns detected.