Scan LLM-generated code before use with CodeShield

Run CodeShield on model-produced code or command suggestions before they reach a user, a repo, or an execution step, so insecure patterns get blocked or warned on first.

Prerequisites

CodeShield integrated into an LLM output pipeline or coding assistant workflow

Installation

No source-backed install or usage instructions could be extracted automatically. Review the upstream project before running this skill in a sensitive workflow.

• Source: https://github.com/meta-llama/PurpleLlama/tree/main/CodeShield

Documentation

• https://github.com/meta-llama/PurpleLlama/tree/main/CodeShield

Source

• Agent Skill Exchange