Sanitize prompts and detect prompt injection before LLM processing.
npx skills add https://github.com/openguardrails/moltguard --skill moltguard
Please help me install this Agent Skill. Skill: moltguard Source: https://github.com/openguardrails/moltguard/tree/main/clawhub-skill Command: npx skills add https://github.com/openguardrails/moltguard --skill moltguard
This Skill protects your AI applications by detecting and blocking prompt injection attacks and sanitizing sensitive data before it's sent to LLMs, ensuring security and privacy.
Install the MoltGuard plugin using the OpenClaw CLI.
High-intent search queries and answers about installing and using this skill.
Prompt injection detection identifies and blocks malicious instructions hidden within user inputs or external content before they reach your LLM. It analyzes prompt content locally to prevent the LLM from being tricked into performing unauthorized actions.
Yes, local prompt sanitization redacts sensitive data like PII, API keys, and credit card numbers directly on your machine. This ensures privacy by preventing sensitive data from being exposed to the LLM provider.
You can securely process user-submitted documents by intercepting the content to sanitize sensitive data and detect malicious prompt injections. This ensures no sensitive data is exposed and the LLM cannot be tricked into unauthorized actions.
Yes, prompt injection detection and local prompt sanitization support various LLM providers. They offer configurable security policies to ensure your AI applications remain secure across different models.
Local prompt sanitization is the process of redacting sensitive data like PII and API keys directly on your machine before sending prompts to an LLM. You need it to ensure data privacy and security when processing external content.
haxlysGenerate and enforce Kubernetes NetworkPolicy, PodSecurityPolicy, and RBAC security policies.
openguardrailsSanitizes sensitive data and detects prompt injection before LLM processing.
intrafindValidates IDs, language codes, and file paths to prevent API security vulnerabilities.
Zhow01Detect and block unsafe prompts and secret leakage in OpenClaw deployments.
