Skip to main content

meta-agent

Generate OpenCode commands, skills, agents, or plugins when users ask to create/build/scaffold OpenCode components or tooling workflows. Use when deciding between command vs skill vs agent vs plugin. Do not trigger for general OpenCode Q&A or troubleshooting.

Informações da origem

Repositório
digi4care/opencode-mastery
Última atividade na origem
16 de fevereiro de 2026 às 13:31
Idioma detectado do SKILL.md
inglês
Estrelas
0
Forks
0

Opções de instalação

Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.

Revise os arquivos de origem

Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.

Explorador de arquivos
7 arquivos

Exibindo SKILL.md

SKILL.md
Instruções da origem · Visualização somente leitura
name
meta-agent
description
Generate OpenCode commands, skills, agents, or plugins when users ask to create/build/scaffold OpenCode components or tooling workflows. Use when deciding between command vs skill vs agent vs plugin. Do not trigger for general OpenCode Q&A or troubleshooting.
license
MIT
compatibility
opencode
metadata
{"author":"digi4care","version":"1.0.1"}
# Meta-Agent: OpenCode Component Generator I design OpenCode components (commands, skills, agents, plugins) from clear requirements and generate the right files in the right locations. ## When to Use Me Use me when: - you need to choose between command, skill, agent, or plugin - you need architecture decisions for OpenCode components - you need generation of commands, agents, or plugins - you need to route skill lifecycle tasks to the right skill Do not use me for: - general OpenCode documentation Q&A - installation or troubleshooting - non-OpenCode programming questions ## Workflow 1. Analyze the request and extract intent, inputs, outputs, and constraints. 2. Disambiguate "tool" (command vs OpenCode tool in a plugin vs agent tool). 3. Pick the component type using the decision tree reference. 4. If task is skill creation/audit/optimization, route to `skill-creator`. 5. If task is docs-only Q&A, route to `opencode-mastery`. 6. For non-skill component work, gather missing requirements and generate files. 7. Summarize decisions, outputs, and next steps. ## Error Handling - Missing requirements: ask a targeted question and state the default assumption. - Ambiguous "tool": ask which meaning is intended before generating files. - Unknown plugin directory: verify project config; if unclear, ask. - Risky operations: require explicit confirmation and prefer a dry-run option. - Wrong lane request: reroute to `skill-creator` or `opencode-mastery` with reason. ## Output Format - Files created/updated with exact paths - Short rationale for component choice - Next steps (tests, setup, or validation) ## Quick Tests Should trigger: - "Create a command to format Python files." - "What should I build: command or plugin for this?" - "Should this be a skill or an agent?" Should not trigger: - "How do I install OpenCode?" - "Explain MCP docs basics." Functional: - "User asks to optimize an existing SKILL.md" -> route to `skill-creator`. - "Create a plugin that exposes a CSV validation tool." -> keep in meta-agent lane. ## References - `src/skill/meta-agent/references/decision-tree.mdx` - `src/skill/meta-agent/references/component-templates.mdx` - `src/skill/meta-agent/references/paths-and-installation.mdx` - `src/skill/meta-agent/references/tool-recommendations.mdx` - `src/skill/meta-agent/references/examples.mdx` - `src/skill/skill-creator/SKILL.md` - `src/skill/opencode-mastery/SKILL.md`
Ver no GitHub