🎁 Get the FREE AI Skills Starter Guide β€” Subscribe β†’
BytesAgainBytesAgain
BytesAgain β€Ί Skills β€Ί Chat Distill
πŸ¦€ ClawHub

Chat Distill

by @mengqiyu9

Distill a person's chat style from exported conversation records and generate replies that mimic their voice. Use when (1) analyzing chat history to extract...

Versionv1.0.0
Downloads299
Stars⭐ 1
#writing#audio#legal#education
TERMINAL
clawhub install chat-distill

πŸ“– About This Skill

name: chat-distill description: > Distill a person's chat style from exported conversation records and generate replies that mimic their voice. Use when (1) analyzing chat history to extract vocabulary, tone, emoji habits, sentence patterns, and personality traits, (2) generating replies in someone's specific chat style, (3) creating a style profile from WeChat, TG, Discord, or text chat exports, (4) asking to analyze chat records or mimic a speaker's tone. Supports .txt, .json, and WeChat chat export formats.

Chat Distill β€” Style Analysis & Mimicry

Workflow

1. Parse β†’ extract messages per speaker from raw export (see references/format-parsers.md) 2. Analyze β†’ build style profile (see references/style-dimensions.md) 3. Report β†’ output analysis report using template in references/output-template.md 4. Mimic β†’ generate replies on demand using the profile

Quick Start

Given a chat export file:

1. Read the file and identify the format (WeChat export, plain text, JSON array, TG export). 2. Normalize into { speaker, text, time? } messages using parsing rules in references/format-parsers.md. 3. Pick the target speaker β€” the one whose style to learn. If multiple speakers exist, ask which one. 4. Run analysis following references/style-dimensions.md. 5. Output the report per references/output-template.md Β§ Analysis Report. 6. When the user asks for a mimicked reply, use the profile + references/output-template.md Β§ Mimic Reply.

Key Principles

  • Show, don't tell: Include concrete examples from the actual chat when reporting style traits.
  • Preserve quirks: Capture tics the speaker doesn't notice β€” repeated filler words, capitalization habits, punctuation style.
  • Respect privacy: Never echo sensitive content (passwords, addresses, financials) from chats into reports. Anonymize if needed.
  • Minimum sample: Require at least 20 messages from the target speaker. If fewer, warn that analysis may be unreliable.

    • πŸ’‘ Examples

    Given a chat export file:

    1. Read the file and identify the format (WeChat export, plain text, JSON array, TG export). 2. Normalize into { speaker, text, time? } messages using parsing rules in references/format-parsers.md. 3. Pick the target speaker β€” the one whose style to learn. If multiple speakers exist, ask which one. 4. Run analysis following references/style-dimensions.md. 5. Output the report per references/output-template.md Β§ Analysis Report. 6. When the user asks for a mimicked reply, use the profile + references/output-template.md Β§ Mimic Reply.