Automate creation and orchestration of AI courtroom agents for Agent Court.
npx skills add https://github.com/Uday9316/Nad-Court --skill agent-court-agents
Please help me install this Agent Skill. Skill: agent-court-agents Source: https://github.com/Uday9316/Nad-Court/tree/main/agents Command: npx skills add https://github.com/Uday9316/Nad-Court --skill agent-court-agents
Agent Court AI agents automate the creation and orchestration of specialized courtroom personas (advocate, defender, and multiple judges) to streamline testing and demonstration of AI-based legal workflows.
Instantiate NadCourt-Advocate to draft an opening argument for a security-vulnerability case using the Agent Court prompts.
High-intent search queries and answers about installing and using this skill.
AI courtroom simulations are automated by instantiating specialized agent roles like advocate, defender, and six judge personas to orchestrate multi-round legal proceedings. This skill standardizes role prompts, decision formats, and JSON rubrics for dynamic argument generation and evaluation.
Standardized JSON rubrics are used to evaluate AI judge decisions by defining scoring criteria and safety checks across all agents. This standardizes data flow and enables consistent comparison of outcomes across multiple simulation rounds.
Opening, counter, and closing arguments for AI legal agents are generated using controlled prompts defined within the advocate and defender role identities. These role definitions ensure consistent argument drafting for various simulated case scenarios.
Multi-round legal cases are simulated by orchestrating federated AI courtroom agents through structured proceedings with evaluation and logging workflows. This setup allows researchers to study prompt design, strategy, and governance by comparing judge evaluations across rounds.
Specific personas are required to run an AI agent court simulation, including a NadCourt-Advocate, NadCourt-Defender, and six distinct judge personas. These identities create consistent roles for advocacy, defense, and evaluation throughout the proceedings.
Predefined AI courtroom roles focus on structured legal workflows with fixed advocate, defender, and judge personas, limiting flexibility for unstructured legal scenarios. The rigid JSON rubrics and role definitions are optimized for standardized data flow and scoring rather than open-ended legal research.
walker-hzxAssemble Claude Code Agent Teams with configurable roles and workflows.
matt-rileyCreate approval-gated plans with milestones, risks, and reviewer roles.
pluginagentmarketplaceSimulate AI model extraction attacks and log defense responses.
dhar174Create and run LangSmith evaluation pipelines for AI agents.
