recipe-agent-team-compound-learning

Capture reusable learnings from agent-team runs into compound-learning artifacts.

2|Updated May 10, 2026
One-click install
npx skills add https://github.com/tae2089/agent-team --skill recipe-agent-team-compound-learning
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: recipe-agent-team-compound-learning
Source: https://github.com/tae2089/agent-team/tree/main/skills/recipe-agent-team-compound-learning
Command: npx skills add https://github.com/tae2089/agent-team --skill recipe-agent-team-compound-learning

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Capture after-action learnings from non-trivial agent-team runs to accelerate future workflows by making patterns reusable and reducing rework.

Core Features & Use Cases

  • Capture run-scoped learnings that summarize evidence from runs, tasks, reviews, and artifacts.
  • Generate a run artifact at _workspace/{run_id}/compound-learning.md and, when knowledge is reusable, propose or create a durable docs solution under docs/solutions/.
  • Link learnings to relevant run components to enable quick reuse in planning and execution.

Quick Start

Summarize the run evidence and generate the run-scoped compound-learning artifact.

Frequently Asked Questions about recipe-agent-team-compound-learning

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

FAQPage Schema
How do I capture reusable learnings from agent-team runs to speed up future workflows?

Gather evidence from run summaries, task records, event logs, and artifacts after non-trivial agent-team runs, then generate a run-scoped compound-learning artifact to accelerate future workflows.

What is a compound-learning artifact and when do I need to generate one?

A compound-learning artifact is a run-scoped markdown file summarizing reusable patterns from non-trivial runs like bug fixes, architecture decisions, or planning iterations, needed when you want to reduce rework and speed up future tasks.

Can I use this to document active implementation or trivial fixes during a run?

No, this is not for active implementation, uncertain outcomes, or trivial fixes; it applies only after meaningful agent-team runs, tasks, reviews, debugging sessions, architecture decisions, and planning iterations are complete.

How do I create durable solution docs from run evidence?

Summarize run evidence from tasks, reviews, and artifacts to generate the run-scoped compound-learning artifact, which can then propose or create a durable solution doc under the docs/solutions/ directory for broader reuse.

What's the best way to link run learnings to planning and execution components?

The best way to link run learnings is to generate the compound-learning artifact at _workspace/{run_id}/compound-learning.md, which automatically connects the summarized evidence to relevant run components for quick reuse in planning and execution.