ShieldCortex provides AI agents with a secure, persistent memory system featuring semantic search, knowledge graphs, and a 6-layer defense pipeline.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install shieldcortex-skill
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install shieldcortex-skill using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is ShieldCortex?
ShieldCortex is a comprehensive memory management framework designed to give AI agents a persistent brain that remains consistent across sessions. By integrating ShieldCortex into your Openclaw Skills, you enable agents to retain complex context, such as architectural decisions and user preferences, without overloading the context window. It goes beyond simple storage by implementing semantic search, automatic memory consolidation, and decay-based forgetting to mimic natural cognitive processes.
Beyond memory, ShieldCortex prioritizes security through a robust defense pipeline. It protects Openclaw Skills from memory poisoning, prompt injection, and accidental credential leaks. With its ability to scan instruction files like SKILL.md and .cursorrules, it ensures that your agent environment remains secure and resilient against malicious inputs or data corruption.
ShieldCortex Use Cases
- Retaining project-specific architectural decisions and coding preferences across multiple sessions.
- Performing semantic searches across historical interactions to retrieve relevant context without keyword matching.
- Protecting AI agents from prompt injection attacks and malicious memory poisoning through a 6-layer defense.
- Scanning local skill files and agent instructions for hidden security threats and credential leaks.
- Automatically extracting knowledge graphs to map entities and relationships from agent interactions.
How ShieldCortex Works
- The system captures memory writes which are immediately passed through a 6-layer defense pipeline to sanitize input and detect threats.
- Validated data is stored in a persistent SQLite database with associated importance levels, categories, and project scopes.
- Semantic search and activation scoring prioritize the most relevant memories for retrieval during active agent sessions.
- The knowledge graph engine extracts entities and triples to build a relational map of stored information.
- Background processes handle memory consolidation and decay, ensuring the most salient information stays accessible while old data fades.
ShieldCortex Setup
First, install the package globally using npm:
npm install -g shieldcortex
To integrate with your Openclaw Skills environment and activate the cortex-memory hook, run the dedicated installation command:
shieldcortex openclaw install
For other environments like Claude Code, VS Code, or Cursor MCP, use the general install command:
shieldcortex install
ShieldCortex Data Schema & Taxonomy
ShieldCortex organizes information using a sophisticated SQLite-backed schema designed for high-performance retrieval within Openclaw Skills.
| Attribute | Description |
|---|---|
| Content & Title | The primary textual data and its identifier for the memory entry. |
| Importance | Ranging from Low to Critical; determines how quickly a memory decays. |
| Categories | Contextual tags such as Architecture, Decisions, Learnings, or Errors. |
| Knowledge Graph | Structured entities and triples representing relationships between data points. |
| Security Metadata | Trust scores, audit logs, and anomaly detection results for every entry. |
ShieldCortex Advanced Features
- Real-time visual dashboard for monitoring agent memory health and security audits via the CLI.
- Multi-agent project scoping to isolate memories between different workspaces or team environments.
- Automated contradiction detection that flags incoming data conflicting with established persistent memories.
- Credential leak detection engine covering over 25 patterns across 11 major cloud providers.
- Skill scanner tool to audit instruction files like .cursorrules and CLAUDE.md for hidden injection threats.
SKILL.md
Loading
Related Openclaw Skills
Seats.aero Award Flight Search
Search and find award flight availability across 24 mileage programs using the seats.aero partner API.
Mermaid Diagrams Skill
The Mermaid Diagrams Skill enables the programmatic generation of visual charts and diagrams directly from text-based Mermaid syntax.
4claw
A specialized, moderated imageboard protocol that allows AI agents to create threads, post replies, and participate in community-driven boards.
lobsterguard
A comprehensive bilingual security auditor and real-time shield providing 68 automated checks and 11 self-healing fixes for agent environments.
watchclaw
An automated health monitoring and recovery watchdog designed to protect OpenClaw gateways from faulty configuration changes.
Browser Audio Capture
A bridge that enables AI agents to capture and process real-time audio streams from any browser tab without platform-specific integrations.
Featured*
