opencode-config-validation

name	opencode-config-validation
description	Validates OpenCode configuration files for correctness and best practices

OpenCode Configuration Validation

This skill provides comprehensive validation rules for OpenCode configurations.

Overview

Use this skill to validate agent, skill, and command configurations before finalizing them. It ensures configurations follow OpenCode standards and will work correctly.

When to Use

Load this skill when:

• Creating new OpenCode configurations
• Validating existing configurations
• Checking for naming convention compliance
• Ensuring configuration completeness

Agent Validation

Required Elements

1. Frontmatter Fields:

   • description: Required, 1-200 characters
   • mode: Required for agents (primary, subagent, or all)

2. Optional Fields:

   • model: Format must be provider/model-id
     • To get a list of valid model IDs, run: opencode models
     • This command shows all available models with their providers
   • temperature: Must be between 0.0 and 1.0
   • tools: Object with boolean values
   • permissions: Object with allow, ask, or deny values
   • maxSteps: Positive integer for iteration limits

3. File Structure:

   • Location: ~/.config/opencode/agent/[name].md or .opencode/agent/[name].md
   • Must have frontmatter section (between --- markers)
   • Should include agent instructions after frontmatter

Naming Rules

• Lowercase letters only
• Hyphens allowed (not at start/end)
• No spaces, underscores, or special characters
• Valid: code-review, otto-analyzer
• Invalid: Code_Review, otto analyzer, review-

Tool Configuration

Valid tool names:

• read, write, edit
• bash, glob, grep
• skill, task, webfetch
• todowrite, todoread
• MCP server tools (format: servername_toolname)

Permission Configuration

permissions:
  edit: allow|ask|deny
  bash:
    "git *": allow
    "npm *": ask
    "*": deny
  webfetch: allow

Skill Validation

Required Elements

1. Directory Structure:

   • Must be in skill/[name]/SKILL.md
   • Directory name must match skill name

2. Frontmatter Fields:

   • name: Required, must match directory
   • description: Required, 1-1024 characters

3. Optional Fields:

   • license: SPDX identifier
   • compatibility: String value
   • metadata: Key-value pairs

Naming Rules

• Must match: ^[a-z0-9]+(-[a-z0-9]+)*$
• No consecutive hyphens (--)
• No leading/trailing hyphens
• Valid: git-workflow, code-review-process
• Invalid: git--workflow, -review, review-

Content Structure

Should include:

• Overview section
• "When to Use" criteria
• Clear instructions
• Examples (optional but recommended)

Command Validation

Required Elements

1. File Location:

   • Must be in command/[name].md
   • File name becomes command name

2. Frontmatter Fields:

   • At least one of: template or description

3. Optional Fields:

   • agent: Which agent handles the command
   • model: Model override
   • subtask: Boolean for subagent invocation

Template Syntax

• $ARGUMENTS: All arguments as string
• $1, $2, etc.: Individual arguments
• !command``: Shell command interpolation
• @filename: File reference

Naming Rules

• Short, memorable names
• Lowercase only
• No special characters
• Should be easy to type

Common Validation Errors

Agent Errors

1. Missing description: Every agent needs a description
2. Invalid mode: Must be primary, subagent, or all
3. Bad tool names: Check spelling of tool names
4. Invalid temperature: Must be 0.0-1.0

Skill Errors

1. Name mismatch: Directory and name field must match
2. Invalid characters: Only lowercase and hyphens
3. Missing SKILL.md: File must be named exactly SKILL.md
4. Description too long: Keep under 1024 characters

Command Errors

1. No template: Commands need content
2. Invalid placeholders: Check $ syntax
3. Bad shell syntax: Use !command not just !command

Validation Process

1. Structure Check:

   • Verify file locations
   • Check directory structure
   • Validate file naming

2. Frontmatter Validation:

   • Parse YAML correctly
   • Check required fields
   • Validate field values

3. Content Validation:

   • Ensure instructions exist
   • Check for placeholder syntax
   • Verify example validity

4. Cross-Reference Check:

   • No duplicate names
   • Valid agent references
   • Existing skill dependencies

Best Practices

1. Descriptions:

   • Be specific about purpose
   • Use action-oriented language
   • Keep concise but informative

2. Naming:

   • Use descriptive names
   • Follow conventions consistently
   • Avoid overly generic names

3. Documentation:

   • Include usage examples
   • Document special behaviors
   • Explain configuration choices

4. Testing:

   • Test configurations before deployment
   • Verify in isolation first
   • Check integration with other components