member-skill-distiller

Distill team member profiles, files, and YouTube subtitles into session-activated member skills.

1.6k|214|Updated Jun 4, 2025
One-click install
npx skills add https://github.com/Jamailar/RedBox --skill member-skill-distiller
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: member-skill-distiller
Source: https://github.com/Jamailar/RedBox/tree/main/desktop/builtin-skills/member-skill-distiller
Command: npx skills add https://github.com/Jamailar/RedBox --skill member-skill-distiller

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Turning a real team member's identity, speaking style, and expertise into a reusable AI persona is hard to do consistently; ad-hoc prompts lose the member's voice, knowledge boundaries, and evidence sources across sessions.

Core Features & Use Cases

  • Persona Compilation: Compiles a member's profile, advisor-bound files, and imported YouTube subtitles into a durable skill package that speaks as the member, not as a generic assistant.
  • Session-Based Activation: Emits skills activated by session metadata rather than message keyword routing, with separated persona, retrieval scope, tool policy, and evidence references for runtime auditing.
  • Bounded Uncertainty Handling: Instructs the generated skill to state uncertainty briefly and keep giving bounded recommendations when evidence is incomplete.
  • Use Case: A content team wants an AI advisor that answers as their senior strategist; this skill distills the strategist's files and subtitle transcripts into a member skill with id, source type, preferred language, persona, system prompt, and evidence summary.

Quick Start

Distill the selected team member's profile, bound files, and YouTube subtitles into an activatable member skill.

Frequently Asked Questions about member-skill-distiller

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I turn a team member's profile into an AI persona skill?

Compile the member's profile, advisor-bound files, and imported YouTube subtitles through the distiller, which emits a skill containing the member id, persona, system prompt, and evidence summary. The generated skill is activated by session metadata rather than keyword routing.

What sources does member skill distillation use?

It prefers advisor-bound files and imported YouTube subtitles over generic workspace context. These sources preserve the member's speaking style, professional stance, and knowledge boundaries with auditable evidence references.

How is a generated member skill activated in a session?

Activation happens through session metadata, not message keyword routing. The skill package separates persona, retrieval scope, and tool policy so the runtime can activate and audit it per session.

What happens when member evidence is incomplete?

The generated skill instructs the member persona to state uncertainty briefly and continue giving bounded recommendations. This keeps responses useful without fabricating knowledge outside the member's evidence base.

Does the distilled persona speak as the member or as an assistant?

The output requirements mandate that the generated skill tells the model to speak as the member, not as a generic assistant. The persona, preferred language, and system prompt are all included in the emitted package.