#decision-making

Q-Learning Swarm Foraging 2026: Multi-Agent RL in Dynamic Grid Environments

Open-source Agent Skill for Claude Code that makes Claude reason with explicit probabilities, base rates, and falsifiers. Distilled from 51 books; first in 18/18 blind passes vs baseline Claude and GPT-5.5, and benchmarked against 14 rival skills and prompt techniques with full cost telemetry (round 5).

Stress-test an idea before you spend months on it: five adversarial AI personas fight it out in parallel, a judge rules GO, PIVOT or KILL with a pre-registered next test, and the verdict stays accountable in your repo as evidence comes in. Claude Code skill, plain markdown, zero dependencies.

Four-role adversarial council - Believer, Skeptic, Investor, Judge - as a portable Agent Skill. Rulings carry an evidence-tier justification, a ranked list of what must be true, and the cheapest test that would falsify them.

Virtual expert council for Claude Code: assemble real experts on any topic, make them debate, with web fact-check and compounding memory. Bilingual, no API keys, fully local.

One Agent, many minds. Turn a single AI Agent into a multi-perspective meeting with real debate, then get a shareable HTML minutes file. Zero extra infrastructure.

Structured Thinking for Claude Code 2026 – Ask, Search, Verify Before You Build

Structured decision-making for Claude Code: a /method command, an auto-triggering methodology-driven-thinking skill, and a red-team-critic subagent over a shared 29-method index. Methods applied explicitly, not named.

Sun Tzu Strategic AI Advisor 2026 - Competitive Analysis & Decision Engine

Middleware to ensure agents memorize decisions

A framework of prompts and personas for running structured multi-expert deliberation sessions with any AI assistant.

Calibrated decision measurement skills for AI agents.

An agent skill that disagrees with you, on purpose, with evidence. Risk first, pre mortem always, verdict changes only on new evidence.

Claude skill: predicts how Elon Musk would think, decide, react and talk in your situation. Built from primary sources, fact-checked through Oct 2026. Unofficial.

A simple, human-readable method for processing unorganized items, shared by humans and agents.