forge-persona

Building a persona from chat logs requires evidence-based extraction of expression style and interaction patterns from text traces. This Skill distills conversational voice and relationship dynamics from chat logs and social media content into a structured persona.md or persona.json profile.

Relationship-aware persona distillation reconstructs how a specific person responds to you across emotional contexts like venting or conflict. It extracts interaction habits from chat logs and social posts to preserve the relationship dynamic rather than creating a generic imitation.

You can correct and validate the generated persona profile through a built-in validation and correction loop. After generation, you provide feedback on accuracy, and the system applies incremental corrections while using a validator tool to ensure structural integrity and evidence coverage.

Persona distillation works with WeChat chat logs using a dedicated wechat_parser.py dependency. It parses chat log data to extract conversational voice, pacing, and punctuation habits for relationship-aware persona reconstruction.

Maintaining local-first privacy during social media parsing relies on this Skill's local-first architecture. It processes chat logs and social media content on your machine using local parser tools, ensuring personal traces remain private during persona distillation.

Persona generation requires evidence-based extraction to ensure the reconstructed voice accurately reflects real wording, rhythm, and punctuation habits. This prevents generic imitation by grounding the persona profile in actual chat log and social media interaction data.