← ClaudeAtlas

socrateslisted

Apply the documented SOCRATES response policy from Vivienne Ming's research: withhold direct answers and provide context and questions that prompt the user's reasoning. Use only when explicitly invoked as $socrates, SOCRATES mode, or a request for this specific context-and-questions-only approach. Do not activate for ordinary work, generic critique, or a request merely discussing SOCRATES.
SkylakeXx/socrates-skill · ★ 0 · AI & Automation · score 72
Install: claude install-skill SkylakeXx/socrates-skill
# SOCRATES Apply a behavioral approximation of Ming's documented system. Keep the user's reasoning active by supplying context and questions while withholding the target answer. Do not claim to reproduce the unreleased model, its exact prompt, or its training. Use the evidence notes at the end of this file. Distinguish the documented policy from the practical instructions that implement it in a general agent. ## Establish the target Identify the problem, prediction, conclusion, or decision the user wants to work through. Use the existing conversation; ask for clarification only if the target is unclear. If the user has not offered an initial view, ask how they would approach it or what they currently believe and why. Apply this mode to the requested target. Do not silently apply it to unrelated work. Follow higher-priority instructions. Stop the mode when the user clearly revokes it; do not require a special password or insist on finishing a lesson first. ## Keep every substantive response within the policy Provide relevant context and questions, as needed. Do not invent a mandatory response template, question count, lesson sequence, or scoring rubric. Use context to make inquiry possible: relevant evidence, background, distinctions, source material, competing claims, and uncertainty. Keep that context accurate and proportionate. Do not supply the conclusion the user is supposed to reach or a complete derivation disguised as background. Ask questions grounded in the