Back to skills
SkillHub ClubShip Full StackFull Stack
agent-selection
Imported from https://github.com/metasaver/metasaver-marketplace.
Packaged view
This page reorganizes the original catalog entry around fit, installability, and workflow context first. The original raw source lives below.
Stars
0
Hot score
74
Updated
March 20, 2026
Overall rating
C2.9
Composite score
2.9
Best-practice grade
B77.6
Install command
npx @skill-hub/cli install metasaver-metasaver-marketplace-agent-selection
Repository
metasaver/metasaver-marketplace
Skill path: plugins/metasaver-core/skills/cross-cutting/agent-selection
Imported from https://github.com/metasaver/metasaver-marketplace.
Open repositoryBest for
Primary workflow: Ship Full Stack.
Technical facets: Full Stack.
Target audience: everyone.
License: Unknown.
Original source
Catalog source: SkillHub Club.
Repository owner: metasaver.
This is still a mirrored public skill entry. Review the repository before installing into production workflows.
What it helps with
- Install agent-selection into Claude Code, Codex CLI, Gemini CLI, or OpenCode workflows
- Review https://github.com/metasaver/metasaver-marketplace before adding agent-selection to shared team environments
- Use agent-selection for development workflows
Works across
Claude CodeCodex CLIGemini CLIOpenCode
Favorites: 0.
Sub-skills: 0.
Aggregator: No.
Original source / Raw SKILL.md
---
name: agent-selection
description: Single source of truth for MetaSaver agent selection and subagent_type mapping
---
# Agent Selection Skill
Complete reference for selecting the correct MetaSaver agents and their `subagent_type` values.
## Core Principle
**When MetaSaver plugin is active, ALWAYS use MetaSaver agents over core Claude Code agents.**
## Core Agent Replacement
**In MetaSaver repos, prefer MetaSaver agents over core Claude Code agents:**
| Core Agent | Use Instead | Why | subagent_type |
| ----------------- | --------------- | ------------------------------ | ------------------------------------------ |
| `Explore` | `code-explorer` | Uses Serena, repomix, memories | `core-claude-plugin:generic:code-explorer` |
| `Plan` | `architect` | Knows MetaSaver patterns | `core-claude-plugin:generic:architect` |
| `general-purpose` | Task-specific | Specialized knowledge | See tables below |
**In MetaSaver repos:**
- `Explore` → use `core-claude-plugin:generic:code-explorer`
- `Plan` → use `core-claude-plugin:generic:architect`
- `general-purpose` → use task-specific MetaSaver agent
---
## Analysis Agents (1)
**Minimal-tool agents** for prompt classification. These specify `tools: TodoWrite` to avoid inheriting MCP tools (~37k tokens each). Results are returned inline in the Task response.
| Task Type | Agent Name | subagent_type |
| -------------------- | ----------------- | ---------------------------------------------- |
| Scope identification | scope-check-agent | `core-claude-plugin:generic:scope-check-agent` |
**Used by:** analysis-phase skill (first phase of /ms, /audit, /build commands)
---
## Generic Agents (15)
High-level workers for orchestration, implementation, and quality.
| Task Type | Agent Name | subagent_type |
| ------------------------ | ---------------------- | --------------------------------------------------- |
| Codebase exploration | code-explorer | `core-claude-plugin:generic:code-explorer` |
| Agent/skill authoring | agent-author | `core-claude-plugin:generic:agent-author` |
| Architecture decisions | architect | `core-claude-plugin:generic:architect` |
| Requirements analysis | business-analyst | `core-claude-plugin:generic:business-analyst` |
| Code implementation | coder | `core-claude-plugin:generic:coder` |
| Backend development | backend-dev | `core-claude-plugin:generic:backend-dev` |
| Testing | tester | `core-claude-plugin:generic:tester` |
| Code review | reviewer | `core-claude-plugin:generic:reviewer` |
| Multi-agent coordination | project-manager | `core-claude-plugin:generic:project-manager` |
| Quality validation | code-quality-validator | `core-claude-plugin:generic:code-quality-validator` |
| Debugging/investigation | root-cause-analyst | `core-claude-plugin:generic:root-cause-analyst` |
| Security analysis | security-engineer | `core-claude-plugin:generic:security-engineer` |
| Performance optimization | performance-engineer | `core-claude-plugin:generic:performance-engineer` |
| DevOps/infrastructure | devops | `core-claude-plugin:generic:devops` |
| Azure DevOps | azure-devops-agent | `core-claude-plugin:generic:azure-devops-agent` |
---
## Domain Agents (12)
Specialized agents for specific technical domains.
### Backend (2)
| Domain | Agent Name | subagent_type |
| ------------------- | ------------------------- | ------------------------------------------------------------- |
| Data Service API | data-service-agent | `core-claude-plugin:domain:backend:data-service-agent` |
| Integration Service | integration-service-agent | `core-claude-plugin:domain:backend:integration-service-agent` |
### Database (1)
| Domain | Agent Name | subagent_type |
| --------------- | --------------------- | ---------------------------------------------------------- |
| Prisma Database | prisma-database-agent | `core-claude-plugin:domain:database:prisma-database-agent` |
### Frontend (5)
| Domain | Agent Name | subagent_type |
| --------------------- | ---------------------- | ----------------------------------------------------------- |
| React App Scaffolding | react-app-agent | `core-claude-plugin:domain:frontend:react-app-agent` |
| React Components | react-component-agent | `core-claude-plugin:domain:frontend:react-component-agent` |
| MFE Host | mfe-host-agent | `core-claude-plugin:domain:frontend:mfe-host-agent` |
| MFE Remote | mfe-remote-agent | `core-claude-plugin:domain:frontend:mfe-remote-agent` |
| shadcn Components | shadcn-component-agent | `core-claude-plugin:domain:frontend:shadcn-component-agent` |
### Testing (3)
| Domain | Agent Name | subagent_type |
| ----------------- | ---------------------- | ---------------------------------------------------------- |
| Unit Tests | unit-test-agent | `core-claude-plugin:domain:testing:unit-test-agent` |
| Integration Tests | integration-test-agent | `core-claude-plugin:domain:testing:integration-test-agent` |
| E2E Tests | e2e-test-agent | `core-claude-plugin:domain:testing:e2e-test-agent` |
### Monorepo (1)
| Domain | Agent Name | subagent_type |
| -------------- | -------------------- | --------------------------------------------------------- |
| Monorepo Setup | monorepo-setup-agent | `core-claude-plugin:domain:monorepo:monorepo-setup-agent` |
---
## Config Agents (28)
Configuration file specialists for auditing and building configs.
### Build Tools (8)
| Config Type | Agent Name | subagent_type |
| -------------- | -------------------- | -------------------------------------------------------------------------- |
| Docker Compose | docker-compose-agent | `core-claude-plugin:config:build-tools:docker-compose-agent` |
| Dockerignore | dockerignore-agent | `core-claude-plugin:config:build-tools:dockerignore-agent` |
| pnpm Workspace | pnpm-workspace-agent | `core-claude-plugin:config:build-tools:pnpm-workspace-configuration-agent` |
| PostCSS | postcss-agent | `core-claude-plugin:config:build-tools:postcss-agent` |
| Tailwind | tailwind-agent | `core-claude-plugin:config:build-tools:tailwind-agent` |
| Turbo | turbo-config-agent | `core-claude-plugin:config:build-tools:turbo-config-agent` |
| Vite | vite-agent | `core-claude-plugin:config:build-tools:vite-agent` |
| Vitest | vitest-agent | `core-claude-plugin:config:build-tools:vitest-agent` |
### Code Quality (3)
| Config Type | Agent Name | subagent_type |
| ------------ | ------------------ | --------------------------------------------------------------------- |
| EditorConfig | editorconfig-agent | `core-claude-plugin:config:code-quality:editorconfig-agent` |
| ESLint | eslint-agent | `core-claude-plugin:config:code-quality:eslint-agent` |
| Prettier | prettier-agent | `core-claude-plugin:config:code-quality:prettier-configuration-agent` |
### Version Control (5)
| Config Type | Agent Name | subagent_type |
| --------------- | --------------------- | ----------------------------------------------------------------- |
| Commitlint | commitlint-agent | `core-claude-plugin:config:version-control:commitlint-agent` |
| Gitattributes | gitattributes-agent | `core-claude-plugin:config:version-control:gitattributes-agent` |
| GitHub Workflow | github-workflow-agent | `core-claude-plugin:config:version-control:github-workflow-agent` |
| Gitignore | gitignore-agent | `core-claude-plugin:config:version-control:gitignore-agent` |
| Husky | husky-agent | `core-claude-plugin:config:version-control:husky-git-hooks-agent` |
### Workspace (12)
| Config Type | Agent Name | subagent_type |
| ----------------------- | ----------------------------- | -------------------------------------------------------------------- |
| CLAUDE.md | claude-md-agent | `core-claude-plugin:config:workspace:claude-md-configuration-agent` |
| .env.example | env-example-agent | `core-claude-plugin:config:workspace:env-example-agent` |
| Monorepo Root Structure | monorepo-root-structure-agent | `core-claude-plugin:config:workspace:monorepo-root-structure-agent` |
| Nodemon | nodemon-agent | `core-claude-plugin:config:workspace:nodemon-agent` |
| .npmrc.template | npmrc-template-agent | `core-claude-plugin:config:workspace:npmrc-template-agent` |
| .nvmrc | nvmrc-agent | `core-claude-plugin:config:workspace:nvmrc-agent` |
| README | readme-agent | `core-claude-plugin:config:workspace:readme-agent` |
| Repomix | repomix-config-agent | `core-claude-plugin:config:workspace:repomix-config-agent` |
| Root package.json | root-package-json-agent | `core-claude-plugin:config:workspace:root-package-json-agent` |
| Scripts | scripts-agent | `core-claude-plugin:config:workspace:scripts-agent` |
| TypeScript | typescript-agent | `core-claude-plugin:config:workspace:typescript-configuration-agent` |
| VS Code | vscode-agent | `core-claude-plugin:config:workspace:vscode-agent` |
---
## Agent Spawning Pattern
Tell agents to READ their own instruction file:
```
[MODE] for [path/scope].
You are [Agent Name].
READ YOUR INSTRUCTIONS at .claude/agents/[category]/[agent-name].md
Follow YOUR rules, invoke YOUR skills, use YOUR output format.
```
**Example:**
```
Task(
subagent_type: "core-claude-plugin:config:build-tools:eslint-agent",
prompt: `
AUDIT MODE for /path/to/eslint.config.js
You are the ESLint Agent.
READ YOUR INSTRUCTIONS at .claude/agents/config/code-quality/eslint-agent.md
Follow YOUR rules, invoke YOUR skill, use YOUR output format.
Report violations and recommendations.
`
)
```
Results are returned inline.
---
## Quick Reference by Task
| I need to... | Use this agent | subagent_type |
| --------------------- | ---------------------- | -------------------------------------------------------------------- |
| Explore codebase | code-explorer | `core-claude-plugin:generic:code-explorer` |
| Design architecture | architect | `core-claude-plugin:generic:architect` |
| Write code | coder | `core-claude-plugin:generic:coder` |
| Write tests | tester | `core-claude-plugin:generic:tester` |
| Review code | reviewer | `core-claude-plugin:generic:reviewer` |
| Coordinate agents | project-manager | `core-claude-plugin:generic:project-manager` |
| Validate quality | code-quality-validator | `core-claude-plugin:generic:code-quality-validator` |
| Debug issues | root-cause-analyst | `core-claude-plugin:generic:root-cause-analyst` |
| Audit ESLint | eslint-agent | `core-claude-plugin:config:code-quality:eslint-agent` |
| Audit TypeScript | typescript-agent | `core-claude-plugin:config:workspace:typescript-configuration-agent` |
| Build React component | react-component-agent | `core-claude-plugin:domain:frontend:react-component-agent` |
| Build API service | data-service-agent | `core-claude-plugin:domain:backend:data-service-agent` |
| Setup Prisma | prisma-database-agent | `core-claude-plugin:domain:database:prisma-database-agent` |
---
## Integration
This skill is referenced by:
- `/ms` - MetaSaver intelligent router
- `/audit` - Configuration and code auditing
- `/build` - Feature and component building
All commands should use this skill as the single source of truth for agent selection.