balanced
The anti-sycophancy engine. Forces multi-perspective analysis before you commit.
Constructive, evidence-based dialogue mode that avoids sycophancy. This skill should be used when the user wants balanced multi-perspective analysis, critical feedback, or rigorous challenge of their ideas. Triggers on "/balanced" or requests for honest/critical/balanced feedback. Supports passive, interactive, tldr, steelman, and decision modes.
What it does
Activates an evidence-based dialogue mode that refuses to be a yes-man. Instead of agreeing with whatever you say, it runs a structured four-move analysis: surface merits, rigorous challenge, alternative framings, and practical synthesis.
Modes
- Full (default) — complete 4-move structured analysis in a single pass
- Interactive — Socratic Q&A, one move at a time, with user input between each
- TL;DR — 3–5 lines: one key fact, one challenge, one action
- Steelman — builds the strongest version of your argument AND the strongest counter-argument
- Decision — tradeoff table with an explicit call, for when analysis is done and you need to decide
When to use
Before committing to a strategy, making a significant decision, or whenever you need honest pushback rather than polite agreement. Run /balanced followed by your statement or question.