agentsclimarketplace

Moe

Skill tiennt235/moe/skill/moe

A team of domain experts as installable agent skills + subagents for Claude Code, Codex, and Pi. Each expert answers from a curated, cited knowledge folder - no RAG, no vector DB, no embeddings.From the repository description

Install
npx -y skills add tiennt235/moe --skill moe

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.0 KB, 498 tokens by cl100k_base, as published. Nobody here has run it

moe — mixture of domain experts

Route a domain question to the right expert(s) on the team and return a cited answer. The team and what each expert knows:

{{ROSTER}}

When to use

Use this for substantive questions that one of the experts above covers. For anything outside the team's domains, answer normally — do not force-fit an expert.

How to answer — /moe ask "<question>"

  1. Select the expert(s) whose description best matches the question. More than one may apply; pick every one that clearly does (usually 1–2). If none fit, say so and stop.
  2. Delegate to each selected expert: {{DELEGATION}}
  3. Synthesize one answer from what the expert(s) return. Preserve every citation (source · section · page). If experts disagree, surface both sides with their citations.
  4. If the knowledge base does not contain the answer, say so plainly — never invent.

Commands

  • /moe ask "<q>" — route → answer (this is the default).
  • /moe route "<q>" — show which expert(s) you would pick and why; do not answer.
  • /moe list — list the team and what each expert knows.
  • /moe add-expert <name> — add a new expert (dev-path/maintainer action; see its command).
  • /moe build — rebuild knowledge + per-host outputs (dev-path/maintainer action).

Creating or rebuilding experts is a maintainer action

A request to create/build/add a new expert (e.g. "build an AWS ML expert") is not a domain question — do not route it to a domain expert and do not try to satisfy it by searching knowledge folders or agent files. Follow the add-expert / build commands: if the moe-expert-builder subagent is present (dev build) delegate to it, otherwise tell the user to use the dev path (clone the repo + Python) and stop. Building needs the moe repo + Python; a plain install cannot do it.

Every expert answers only from its own knowledge folder and cites it. Routing is your judgment over the descriptions above — there is no vector search and no embeddings.

What ships with it: 5 files

3.5 KB alongside SKILL.md

commands/

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.