Shield Cn

CN本地化安全防护技能。针对中文场景设计的安全加固方案，防御prompt注入、数据泄露、凭证泄漏，支持钉钉/飞书/微信等国内平台的安全检测。

This skill appears to do what it claims — local scanning and interactive guarding for Chinese scenarios — but it operates on files you point it at and writes logs/reports to your home directory. Before installing/running: 1) inspect the scripts (they're small and included) and run them in a test/sandbox workspace first; 2) avoid scanning your entire filesystem or running as root (the auditor will happily walk whatever directory you give it); 3) be aware reports may contain short redacted snippets of matched secrets (first ~20 chars + '***') and those will be stored in reports/logs — remove or secure reports if needed; 4) verify the upstream repository/author if you require provenance; and 5) if you want to integrate into production agents, review/adjust default rules, whitelists, and log storage locations to match your security policy.

Type: OpenClaw Skill Name: shield-cn Version: 2284.0.0 The 'shield-cn' skill is a localized security toolkit designed to protect AI agents in Chinese-language environments. It provides scripts for auditing local workspaces for exposed credentials (security-audit.py) and monitoring user inputs for prompt injection or phishing attempts (shield-guard.py). The implementation relies on standard Python libraries for local file scanning and regex-based detection, with no evidence of data exfiltration, network communication, or malicious intent.