forge-self

What problem does it solve?

You lack a reliable way to see your own patterns—how you decide, what you value, and where your blind spots are—so your choices feel cloudy or inconsistent.

Core Features & Use Cases

• Guided self-distillation: Runs a multi-round intake conversation to capture your identity, decision style, values, and blind spots without judgment or beautifying your reality.
• Optional cross-source evidence analysis: If you provide materials (e.g., WeChat exports, diary/notes, social content), it parses and cross-validates what you say with what you do, preserving contradictions for your confirmation.
• A calibrated “persona base” for reflection: Produces a structured persona.md (and persona.json) that you can later use as input for private decision support (via /use-self).
• Validation + correction loop: Shows a summary, asks what feels “too off” or “too beautified,” records user corrections into a correction layer, and runs an automated validator for structural and evidence coverage.

Quick Start

Tell the user to run /forge-self and follow the four intake phases (conversation first, then optional material import) to generate your persona base.