# titanwings/distilly

Distilly — Distill how they think into reusable Skills for any Agent or Bot. Formerly Colleague Skill（原同事 Skill）.

Repository: https://github.com/titanwings/distilly
Canonical: https://ross.abutalabs.com/products/distilly
Language: Python
License: MIT
License Family: permissive
Topics: agent-skills, ai-agent, claude-code, codex, hermes-agent, knowledge-distillation, meta-skill, openclaw, agentic-ai, ai-persona, codex-skills, deepseek-harness, digital-human, llm-agent, ai-agents, ai-assistants, claude-skills, opencode, dsh-plugin
Last push: 2026-08-25T07:18:07+00:00

## Health v2 (maintenance only)
Score: 63/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 45, longevity 11
- inputs: {"age_days": 156, "days_push": 8, "days_rel": 156, "gap_med": null, "n_releases_24m": 1}
- flags: prerelease_only, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 24016, forks 2111 (observed 2026-08-28T04:11:35.498171+00:00)

## What it is
Distilly is a Python tool that distills a person's experience, judgment, voice, and working style from messages, documents, interviews, and public sources into a reusable, source-grounded Person Profile packaged as an Agent Skill. These profiles can be installed and invoked by AI agents and compatible bots such as Claude Code, Codex, and other agent harnesses.

## Use cases
- capture a departing colleague's decision-making playbook as an agent skill
- create an AI persona of a favorite author or public figure to ask questions
- preserve how a friend or family member communicates for a chatbot
- model my own working style so agents can assist me consistently
- turn interview transcripts into a reusable person profile for agents
- build a digital human persona from chat logs and documents

## When to choose
- you want a portable, source-grounded person profile for AI agents rather than a freeform chatbot clone
- you use agent hosts that support the AgentSkills standard like Claude Code or Codex
- you need to preserve institutional knowledge from departing teammates
- you want to model personas from your own provided materials with provenance

## When to avoid
- you expect an exact clone or faithful simulation of a real person's mind
- you need a hosted SaaS with no local setup
- you lack source material about the person you want to model
- you need real-time conversational memory rather than a static distilled profile

## Facets
- artifact type: library
- maturity: active
- function: agent-framework, prompt-engineering, llm-inference, nlp, chatbot
- domain: artificial-intelligence, large-language-models, chatbots, developer-tools
- platform: python, cli, cross-platform
- tags: agent-skills, person-profile, knowledge-distillation, ai-persona, claude-skills, digital-human, meta-skill, persona-modeling, ai-agents

## Member repositories
- titanwings/distilly (main) score 63

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:35.498171+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:56:18.146033+00:00, confidence not recorded.
  - readme: https://github.com/titanwings/distilly (fetched 2026-08-28T04:11:35.498171+00:00, sha 2058e2186361)
- Data as of 2026-08-30T08:39:29.467469+00:00.