Aegis Shield is a security screening tool designed to prevent prompt injection and data exfiltration by sanitizing untrusted text before it interacts with your AI agent memory or logic.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install aegis-shield
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 aegis-shield 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 Aegis Shield?
Aegis Shield serves as a critical security firewall for AI agents, specifically built to handle untrusted input from web fetches, emails, or social media content. It acts as a gatekeeper within the Openclaw Skills ecosystem, ensuring that malicious prompts or exfiltration attempts are neutralized before they can compromise your system.
By implementing a scan-lint-accept-or-quarantine workflow, it provides a safe environment for AI memory updates and tool interactions. This skill is essential for developers who need to process third-party content without exposing their agent's underlying instructions or sensitive internal data to external manipulation.
Aegis Shield Use Cases
- Scanning web content or email summaries for hidden prompt injection attacks.
- Sanitizing untrusted external text before writing declarative entries to long-term memory.
- Preventing sensitive data exfiltration by screening content before it is processed by other LLM tools.
- Automating the quarantine of suspicious inputs based on configurable severity scores and linting rules.
How Aegis Shield Works
- Receive a chunk of untrusted text from an external source like a website or user message.
- Run a local scan to evaluate the content for injection, tool-abuse patterns, and exfiltration attempts.
- Assign a severity score and category to the input based on the detected risk levels.
- Execute a linting process to ensure the data follows internal safety protocols, such as checking for exposed secrets or keys.
- Accept and append the sanitized data to memory or move the original content to a quarantine directory if the risk exceeds the defined threshold.
Aegis Shield Setup
To begin using this security layer within your Openclaw Skills environment, ensure you have the required scripts available in your local repository.
# Navigate to your agent directory
# Use the bundled script for safe memory operations
node scripts/openclaw-safe-memory-append.js \
--source "web_fetch:https://example.com" \
--tags "ops,security" \
--allowIf medium \
--text "<untrusted content>"
Aegis Shield Data Schema & Taxonomy
The skill organizes its output and tracking through a structured JSON response and a specific directory hierarchy for safe storage.
| Attribute | Description |
|---|---|
status |
The final state of the operation: accepted or quarantined |
source |
The origin of the untrusted text for traceability |
severity |
The risk level identified during the scanning process |
written_to |
File path where sanitized memory was appended |
quarantine_to |
File path where hostile or suspicious content was isolated |
Aegis Shield Advanced Features
- Safe memory append workflow that prevents direct writes to sensitive files, ensuring integrity.
- Configurable severity thresholds using the
--allowIfflag to customize security strictness. - Local-only scanning logic to ensure untrusted text is analyzed before being sent to an LLM provider.
- Automatic tagging and source attribution for all declarative memory entries to maintain a clean data lineage in Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Smart Spawn API
A dynamic routing API that identifies the optimal AI model for specific tasks based on performance, cost, and complexity.
Clawracle Oracle Resolver
An AI agent skill that enables participation in the Clawracle decentralized oracle network on Monad to earn tokens for data resolution.
strawpoll-cli
A comprehensive command-line interface for the StrawPoll API v3 that enables the creation and management of multiple-choice, meeting, and ranking polls.
LaMetric CLI
A comprehensive command-line tool for sending notifications and managing LaMetric TIME/SKY smart displays.
Signet API
A specialized trust scoring platform that provides autonomous AI agents with a verifiable identity and reputation score to enable secure transactions.
Arbitrage Scanner
A real-time monitoring tool designed to identify and calculate profitable price discrepancies across decentralized exchanges on multiple blockchains.
Crypto Price Checker
A high-performance crypto market tracker for real-time price, volume, and 24h change data.
Featured*
