Evaluate decisions through structured multi-perspective debate with Toulmin arguments.
npx skills add https://github.com/jscott3201/ai-agent-skills --skill debate-jscott3201
Please help me install this Agent Skill. Skill: debate Source: https://github.com/jscott3201/ai-agent-skills/tree/main/skills/debate Command: npx skills add https://github.com/jscott3201/ai-agent-skills --skill debate-jscott3201
This Skill structures high-stakes or ambiguous decisions into a repeatable, multi-perspective debate process that surfaces trade-offs, hidden risks, and residual disagreements so teams reach better-justified rulings.
Use the debate skill to evaluate whether to build an in-house vector index or adopt a managed service and produce a synthesized ruling with documented residual disagreements.
High-intent search queries and answers about installing and using this skill.
A multi-perspective debate structures the decision into three phases: independent perspective generation, 2-3 rounds of adversarial exchange, and synthesis with scoring and rulings to surface trade-offs and hidden risks.
The Toulmin model structures arguments with five elements: claim, grounds, warrant, qualifier, and rebuttal. This framework improves evidence quality by requiring every perspective to explicitly state its supporting evidence and potential counterarguments.
Resolve conflicting perspectives through adversarial exchange rounds followed by synthesis. The process produces a justified ruling, scores the arguments, and documents residual disagreements so teams understand remaining trade-offs.
Yes, SeleneDB integration allows optional session recall of prior debates and graph writes to persist perspectives, rulings, and residual disagreements for future reference and analysis.
A multi-perspective debate analysis requires 3-5 distinct perspectives, including a Devil's Advocate, to effectively evaluate technical, product, or strategic choices and surface hidden risks.
Use a structured debate process for high-stakes or ambiguous product trade-offs where you need to surface hidden risks, evaluate multiple perspectives, and reach better-justified rulings through adversarial exchange.
nathan-a-kingAnalyze arguments using Claim-Evidence-Warrant and Toulmin frameworks to identify gaps.
jscott3201Capture lightweight annotation notes for project and session graph nodes.
lleontor705Coordinate adversarial debate rounds and synthesize a final recommendation.
flosrnOrchestrate structured adversarial debates between AI councillors to evaluate ideas.
