socratic

The Socratic skill is a systematic, dialogue-based framework designed for AI assistants to facilitate critical thinking. Unlike standard instructional responses, this agent skill functions as a thought partner, guiding users through structured layers of inquiry to examine their own ideas, proposals, or core beliefs. By shifting from telling to asking, the agent helps users identify gaps in logic, evaluate the validity of their assumptions, and anticipate counter-arguments, making it an essential tool for project managers, software engineers, and researchers who need to validate strategies before implementation.

The skill processes complex problems by breaking them down into five distinct phases: Clarification, Assumption Probing, Evidence Verification, Implication Analysis, and Perspective Synthesis. It is particularly effective for high-stakes decision-making where cognitive biases like confirmation bias or overgeneralization might otherwise skew the outcome. Whether you are debugging a technical architectural choice, debating a product feature, or exploring ethical dilemmas, this skill ensures that all angles of a position are thoroughly stress-tested through adversarial inquiry and logical scrutiny.

• Employs a multi-layered questioning methodology: Clarifying, Probing Assumptions, Examining Evidence, Exploring Implications, and Considering Alternatives.

• Enables 'Steel Manning' of opposing arguments, allowing users to build the strongest possible version of a counter-argument to test their own position's resilience.

• Facilitates logical reasoning checks, helping detect fallacies such as non-sequiturs or circular reasoning in project proposals or strategic plans.

• Promotes a culture of continuous learning by encouraging 'I don't know' as a valid, honest, and high-value answer.

• Provides a structured output format including a summary of refined positions and remaining open questions to ensure actionable takeaways.

• Use this skill when you need to challenge a team's consensus, validate a new business proposal, or conduct a deep dive into complex technical requirements.

• Input your primary claim or belief as a prompt, and the agent will initiate the sequence, starting with boundary-setting questions to define the scope of the discussion.

• Follow the prompts to provide context; the more specific your examples, the better the agent can probe for hidden premises and faulty logic.

• The tool is designed for depth, not speed; expect to engage in a multi-turn interaction to arrive at a truly synthesized and robust conclusion.

• Practical constraint: It is not a direct answer engine; it expects the user to actively participate in the inquiry process rather than passively consuming an generated result.