
from agent-almanac24
Create a reusable multi-agent team definition (YAML + human docs) that standardizes coordination, task decomposition, and activation for Claude Code-style workf
Create Team provides a tested template and step-by-step process to author a machine-readable and human-friendly team definition that coordinates multiple agents to solve multi-stage or multi-expert problems. It guides authors through purpose, lead and member selection, coordination patterns (hub-and-spoke, sequential, parallel, timeboxed, adaptive), task decomposition, and an embedded CONFIG YAML block so teams can be activated programmatically in Claude Code environments. The skill also covers registry updates, README automation, and translation scaffolding.
Use Create Team when a task requires multiple specialized agents (e.g., security audit + code review + architecture review), when work benefits from formal handoffs or synthesis, or when you want a repeatable, deployable team artifact that tooling can spawn. It's appropriate for code review pipelines, compliance validation, multi-stage data pipelines, and recurring coordination workflows.
Primarily designed for Claude Code-style agent runtimes and registries that support spawning subagents from CONFIG blocks; applicable where agents are represented as registry IDs and tooling can parse the CONFIG YAML.
This is a Chinese-localized (zh-CN) skill for creating multi-agent team definition files (YAML + human docs) for the agent-almanac framework. It has no bundled scripts — it's purely instructional, guiding the agent through 11 steps to define teams, write config, register them, and generate translations. The SKILL.md is well-structured with clear steps, failure handling, and a validation checklist, but the content is in Chinese, which limits accessibility for English-speaking users. No security concerns detected.
Well-structured instructional skill with no scripts. Thorough step-by-step process with failure handling at each step. Nice validation checklist. The i18n/zh-CN path suggests this is a translated variant — the English original likely scores higher on breadth/usefulness. No security issues: no remote code execution, no credential handling, no destructive commands. Minor concern: grep commands and npm scripts assume specific repo structure that may not exist elsewhere.
Markov Chain Modeling
Build and analyze discrete or continuous Markov chains: transition matrices, state classification, stationary distributions, and mean first-passage times.
Build Parameterized Report
Generate multiple customized reports from one Quarto or R Markdown template by supplying parameter sets and automating batch renders.
Create Team
Author a multi-agent team composition file (coordination pattern, members, tasks, CONFIG block) for the Agent Almanac teams registry.
Generate Tour Report
Create a Quarto-based, self-contained travel report (HTML/PDF) with embedded maps, daily itineraries, logistics tables, and accommodation/transport details for
Track ML Experiments (MLflow)
Set up MLflow experiment tracking: server, autologging, artifact storage, run comparison and lifecycle management for reproducible ML workflows.
Annotate Source Files
Insert PUT-style workflow annotations into source files (30+ languages) to generate structured workflows and Mermaid diagrams using putior.
Cut Gemstone
Practical, step-by-step guidance for cutting gemstones (cabochon or faceted): assessment, orientation, dopping, grit sequence, and basic faceting angles.
Develop HPLC Method
Design and optimize an HPLC method (RP, NP, IC, HILIC) including column selection, mobile phase, gradient, and detection for robust analytical separations.
Bonsai Cultivation (de)
Comprehensive German-language guide for bonsai care: species selection, pruning, wiring, repotting, seasonal maintenance and contemplative pre-check rituals.
Plan Release Cycle
Create a structured release plan with milestones, feature-freeze rules, RC process, go/no-go checklist and rollback strategy to safely ship software releases.