Skill Detail
Filter prompts and model outputs for injection, secrets, toxicity, and policy risks with LLM Guard
Screen prompts and responses with input and output scanners before an LLM interaction reaches production users or downstream systems.
Security & VerificationMulti-Framework
Security & Verification
Multi-Framework
Security Reviewed
⭐ 2.8k GitHub stars
COPY SKILL INSTRUCTIONS (OPTIONAL)
npx skills add agentskillexchange/skills --skill filter-prompts-and-model-outputs-for-injection-secrets-toxicity-and-policy-risks-with-llm-guard
Uses the third-party skills CLI, not an ASE-owned installer. Check your agent’s compatibility. This copies instructions; complete the upstream tool setup below separately.
At a glance
Tools required
Python 3.9+, application or agent code that can wrap LLM input and output handling
Install & setup
Install with
pip install llm-guard, choose the input and output scanners that match your risk profile, and wrap those checks around prompt submission and response handling in your LLM workflow.Author
Protect AI
Publisher
Organization
Last updated
Apr 16, 2026
Quick brief
Use LLM Guard when the immediate job is to scan or sanitize prompts and model outputs for injection attempts, secret leakage, toxicity, or related policy risks. It is most useful as a preflight and postflight safety layer wrapped around LLM calls in applications or agent pipelines. The boundary is prompt and output filtering, not a full security platform, observability stack, or general-purpose agent framework.