Agent Fcot

Falsification Chain of Thought — post-hoc verification of AI judgments using falsificationism to counter sycophancy bias

# FCoT — Verify AI judgments before you trust them **Falsification Chain of Thought** — a Claude Code skill that counters sycophancy bias by systematically attempting to disprove conclusions. ## When to use When the AI says: - "Yes, that's a good approach" - "This refactor is safe" - "Package A is the better choice" Run `/fcot`. It forces the model to: 1. List real counter-arguments 2. Pre-declare specific conditions that would dismiss each one 3. Verify those conditions 4. Revise the judgment if any counter-argument survives ## When NOT to use - **Simple factual questions** — "What ve

When to use

Community skill by o3co. Install from: https://github.com/o3co/agent-fcot