← 返回 Skills 市场
ShieldCortex
作者
jarvis-drakon
· GitHub ↗
· v4.54.15
· MIT-0
6958
总下载
2
收藏
0
当前安装
164
版本数
在 OpenClaw 中安装
/install shieldcortex
功能描述
Memory and defence for AI agents: semantic recall, knowledge graph and decay, plus a memory firewall that scans and enforces against prompt injection, credential leaks and poisoning.
安全使用建议
Install only if you want durable local memory and active memory-write protection. Review the auto-memory, proactive recall, cloud sync, audit logging, self-heal, and binaryPath settings before enabling integrations, and periodically inspect or delete stored memories in ~/.shieldcortex if you handle secrets or confidential work.
能力评估
Credentials
Reading agent configs, rule files, .env files for scanning, and session transcripts is sensitive but coherent for a memory firewall; writes are mostly contained to ~/.shieldcortex, model cache, OpenClaw extension paths, and MCP registration files.
Install Mechanism
Setup is described as user-initiated with confirmations, while postinstall refresh and OpenClaw bootstrap self-heal are bounded and disclosed; the self-heal can copy hook files and delete two legacy hook directories without a prompt.
Instruction Scope
The main SKILL.md broadly discloses transcript capture, persistent memory, hook behavior, MCP config access, local audit previews, active blocking, and optional cloud sync; the bundled HOOK.md is less clear because it says auto-memory is enabled by default while code requires openclawAutoMemory === true.
Persistence & Privilege
The skill persists memories verbatim in a local SQLite database, uses global memory scope in OpenClaw save paths, and writes local audit previews, so users should treat it as durable local data retention even though cloud upload is opt-in.
Purpose & Capability
The skill's memory, scanning, active interception, local audit, and optional cloud sync capabilities fit its stated memory-and-defence purpose, though they are high-impact capabilities.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install shieldcortex - 安装完成后，直接呼叫该 Skill 的名称或使用
/shieldcortex触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v4.54.15
Sync from npm publish v4.54.15
v4.54.14
Sync from npm publish v4.54.14
v4.54.13
Sync from npm publish v4.54.13
v4.54.12
Sync from npm publish v4.54.12
v4.54.11
Sync from npm publish v4.54.11
v4.54.10
Sync from npm publish v4.54.10
v4.54.9
Sync from npm publish v4.54.9
v4.54.8
Sync from npm publish v4.54.8
v4.54.7
Sync from npm publish v4.54.7
v4.54.5
Sync from npm publish v4.54.5
v4.54.4
Sync from npm publish v4.54.4
v4.54.3
Sync from npm publish v4.54.3
v4.54.2
Sync from npm publish v4.54.2
v4.54.1
Sync from npm publish v4.54.1
v4.54.0
Sync from npm publish v4.54.0
v4.53.0
Sync from npm publish v4.53.0
v4.52.3
Sync from npm publish v4.52.3
v4.52.2
Sync from npm publish v4.52.2
v4.52.1
Sync from npm publish v4.52.1
v4.52.0
Sync from npm publish v4.52.0
元数据
常见问题
ShieldCortex 是什么？
Memory and defence for AI agents: semantic recall, knowledge graph and decay, plus a memory firewall that scans and enforces against prompt injection, credential leaks and poisoning. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 6958 次。
如何安装 ShieldCortex？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install shieldcortex」即可一键安装，无需额外配置。
ShieldCortex 是免费的吗？
是的，ShieldCortex 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
ShieldCortex 支持哪些平台？
ShieldCortex 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 ShieldCortex？
由 jarvis-drakon（@jarvis-drakon）开发并维护，当前版本 v4.54.15。
推荐 Skills