Claude 持久化记忆系统初始化

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:27 PM UTC
• Version:
  1.0.0

Overview

This skill transparently creates local Claude memory files for a project and does not show hidden network, credential, or destructive behavior.

Install or run this only in a repository you want indexed. Review the generated `.claude/memory` files and `CLAUDE.md` before committing them or relying on them, because they may contain private project structure details and will guide future Claude sessions. Use `--force` carefully because it can replace existing generated memory content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The skill performs broad repository modifications by creating `.claude/memory/*` files and creating or rewriting `CLAUDE.md`, and it also supports forced overwrite of existing memory files. Although the body describes what it will do, it does not present a clear up-front warning/consent step before scanning and writing, which can surprise users and cause unintended repository changes.

Natural-Language Policy Violations

Medium

• Confidence: 81% confidence
• Finding: The skill defaults generated comments and descriptions to Chinese unless the user specifies `--lang en`. In environments with language-choice or documentation policy requirements, silently choosing a non-user-requested language can create compliance, usability, and review problems, especially because the generated artifacts are committed project files.

VirusTotal

66/66 vendors flagged this skill as clean.