Distilly
Distilly is an open-source project hosted on GitHub under the repository titanwings/distilly. Formerly known as Colleague Skill (原同事 Skill), it is positioned in the AI Agents category. The project's stated purpose is to distill how people think into reusable Skills that can be used by any agent or bot. This positions Distilly as a knowledge-capture and skill-packaging tool for AI-driven workflows. At a high level, Distilly aims to take the reasoning, decision-making patterns, or problem-solving approaches of an individual or an AI system and turn them into portable, reusable Skills. These Skills can then be applied across different agents and bots, reducing the need to recreate prompts or workflows from scratch. The project appears to target teams and individuals who want to preserve expertise and make it available to AI systems in a structured way. Because Distilly is open source, it is freely available for inspection, modification, and integration. The repository is publicly accessible on GitHub, and the project's former name suggests it evolved from an earlier concept focused on colleague skills. While detailed usage instructions are not included in the available summary, the project's focus on reusable Skills aligns with broader efforts to make AI agents more maintainable and knowledge-driven.
- Category
- AI Agents
- Pricing
- Open Source
Problem solved
Distilly addresses the challenge of capturing and reusing expertise in AI agent workflows. Instead of letting valuable reasoning remain in individual prompts, documents, or people's heads, Distilly provides a way to distill that thinking into reusable Skills. This makes it easier for agents and bots to consistently apply proven approaches and for teams to share knowledge across different AI systems.
Use cases
- Capturing expert decision-making processes as reusable agent Skills
- Converting successful prompt or reasoning patterns into portable Skills
- Building a shared Skill library for multiple bots or agents
- Onboarding new AI agents with distilled organizational knowledge
- Standardizing agent behavior across teams by reusing Skills
- Preserving institutional knowledge from experienced colleagues in an agent-compatible format
Audience
- Developers
- ML Engineers
- Data Scientists
- Enterprise
- Startups
Pros
- Open source and free to use
- Focuses on reusable Skills for agents and bots
- Designed to capture and preserve expertise
- Publicly available on GitHub
- Aligns with modern AI agent workflows
Cons
- Detailed usage documentation is not available in the provided summary
- Project appears early-stage or niche
- No clear release version or installation details in the available information
- May require custom integration for specific agent frameworks
- Name change from Colleague Skill may cause confusion
Official website: https://github.com/titanwings/distilly
Canonical LambdaBase page: https://www.lambdabase.com/tools/distilly