agentleFS
pr-pmpr-pm/prpm/.claude/skills/creating-opencode-plugins/SKILL.md

creating-opencode-plugins

Skill124 starsChanged 8 months agoMITView on GitHub

Use when creating OpenCode plugins that hook into command, file, LSP, message, permission, server, session, todo, tool, or TUI events - provides plugin structure, event API specifications, and implementation patterns for JavaScript/TypeScript event-driven modules

  • Reads credentials

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When porting Claude Code hook behavior to OpenCode plugins, use these event mappings:

    line 320
  2. Distinctive:

    When porting Claude hooks to OpenCode plugins, you'll need to rewrite the logic using the OpenCode plugin API.

    line 379

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Creating OpenCode Plugins

Overview

OpenCode plugins are JavaScript/TypeScript modules that hook into 25+ events across the OpenCode AI assistant lifecycle. Plugins export an async function receiving context (project, client, $, directory, worktree) and return an event handler.

When to Use

Create an OpenCode plugin when:

  • Intercepting file operations (prevent sharing .env files)
  • Monitoring command execution (notifications, logging)
  • Processing LSP diagnostics (custom error handling)
  • Managing permissions (auto-approve trusted operations)
  • Reacting to session lifecycle (cleanup, initialization)
  • Extending tool capabilities (custom tool registration)
  • Enhancing TUI interactions (custom prompts, toasts)

Don't create for:

  • Simple prompt instructions (use agents instead)
  • One-time scripts (use bash tools)
  • Static configuration (use settings files)

Quick Reference

Plugin Structure

export const MyPlugin = async (context) => {
  // context: { project, client, $, directory, worktree }

  return {
    event: async ({ event }) => {
      // event: { type: 'event.name', data: {...} }

      switch(event.type) {
        case 'file.edited':
          // Handle file edits
          break;
        case 'tool.execute.before':
          // Pre-process tool execution
          break;
      }
    }
  };
};

Event Categories

CategoryEventsUse Cases
commandcommand.executedTrack command history, notifications
filefile.edited, file.watcher.updatedFile validation, auto-formatting
installationinstallation.updatedDependency tracking
lsplsp.client.diagnostics, lsp.updatedCustom error handling
messagemessage.*.updated/removedMessage filtering, logging
permissionpermission.replied/updatedPermission policies
serverserver.connectedConnection monitoring
sessionsession.created/deleted/error/idle/status/updated/compacted/diffSession management
todotodo.updatedTodo synchronization
tooltool.execute.before/afterTool interception, augmentation
tuitui.prompt.append, tui.command.execute, tui.toast.showUI customization

Plugin Manifest (package.json or separate config)

{
  "name": "env-protection",
  "description": "Prevents sharing .env files",
  "version": "1.0.0",
  "author": "Security Team",
  "plugin": {
    "file": "plugin.js",
    "location": "global"
  },
  "hooks": {
    "file": ["file.edited"],
    "permission": ["permission.replied"]
  }
}

Implementation

Complete Example: Environment File Protection

// .opencode/plugin/env-protection.js

export const EnvProtectionPlugin = async ({ project, client }) => {
  const sensitivePatterns = [
    /\.env$/,
    /\.env\..+$/,
    /credentials\.json$/,
    /\.secret$/,
  ];

  const isSensitiveFile = (filePath) => {
    return sensitivePatterns.some(pattern => pattern.test(filePath));
  };

  return {
    event: async ({ event }) => {
      switch (event.type) {
        case 'file.edited': {
          const { path } = event.data;

          if (isSensitiveFile(path)) {
            console.warn(`⚠️  Sensitive file edited: ${path}`);
            console.warn('This file should not be shared or committed.');
          }
          break;
        }

        case 'permission.replied': {
          const { action, target, decision } = event.data;

          // Block read/share operations on sensitive files
          if ((action === 'read' || action === 'share') &&
              isSensitiveFile(target) &&
              decision === 'allow') {

            console.error(`🚫 Blocked ${action} operation on sensitive file: ${target}`);

            // Override permission decision
            return {
              override: true,
              decision: 'deny',
              reason: 'Sensitive file protection policy'
            };
          }
          break;
        }
      }
    }
  };
};

Example: Command Execution Notifications

// .opencode/plugin/notify.js

export const NotifyPlugin = async ({ project, $ }) => {
  let commandStartTime = null;

  return {
    event: async ({ event }) => {
      switch (event.type) {
        case 'command.executed': {
          const { command, args, status } = event.data;
          commandStartTime = Date.now();

          console.log(`▶️  Executing: ${command} ${args.join(' ')}`);
          break;
        }

        case 'tool.execute.after': {
          const { tool, duration, success } = event.data;

          if (duration > 5000) {
            // Notify for long-running operations
            await $`osascript -e 'display notification "Completed in ${duration}ms" with title "${tool}"'`;
          }

          console.log(`✅ ${tool} completed in ${duration}ms`);
          break;
        }
      }
    }
  };
};

Example: Custom Tool Registration

// .opencode/plugin/custom-tools.js

export const CustomToolsPlugin = async ({ client }) => {
  // Register custom tool on initialization
  await client.registerTool({
    name: 'lint',
    description: 'Run linter on current file with auto-fix option',
    parameters: {
      type: 'object',
      properties: {
        fix: {
          type: 'boolean',
          description: 'Auto-fix issues'
        }
      }
    },
    handler: async ({ fix }) => {
      const result = await $`eslint ${fix ? '--fix' : ''} .`;
      return {
        output: result.stdout,
        errors: result.stderr
      };
    }
  });

  return {
    event: async ({ event }) => {
      // Monitor tool usage
      if (event.type === 'tool.execute.before') {
        console.log(`🔧 Tool: ${event.data.tool}`);
      }
    }
  };
};

Installation Locations

LocationPathScopeUse Case
Global~/.config/opencode/plugin/All projectsSecurity policies, global utilities
Project.opencode/plugin/Current projectProject-specific hooks, validators

Common Mistakes

MistakeWhy It FailsFix
Synchronous event handlerBlocks event loopUse async handlers
Missing error handlingPlugin crashes on errorWrap in try/catch
Heavy computation in handlerSlows down operationsDefer to background process
Mutating event data directlyCauses side effectsReturn override object
Not checking event typeHandles wrong eventsUse switch/case on event.type
Forgetting context destructuringMissing key utilitiesDestructure { project, client, $, directory, worktree }

Event Data Structures

// File Events
interface FileEditedEvent {
  type: 'file.edited';
  data: {
    path: string;
    content: string;
    timestamp: number;
  };
}

// Tool Events
interface ToolExecuteBeforeEvent {
  type: 'tool.execute.before';
  data: {
    tool: string;
    args: Record<string, any>;
    user: string;
  };
}

interface ToolExecuteAfterEvent {
  type: 'tool.execute.after';
  data: {
    tool: string;
    duration: number;
    success: boolean;
    output?: any;
    error?: string;
  };
}

// Permission Events
interface PermissionRepliedEvent {
  type: 'permission.replied';
  data: {
    action: 'read' | 'write' | 'execute' | 'share';
    target: string;
    decision: 'allow' | 'deny';
  };
}

Testing Plugins

// Test plugin locally before installation
import { EnvProtectionPlugin } from './env-protection.js';

const mockContext = {
  project: { root: '/test/project' },
  client: {},
  $: async (cmd) => ({ stdout: '', stderr: '' }),
  directory: '/test/project',
  worktree: null
};

const plugin = await EnvProtectionPlugin(mockContext);

// Simulate event
await plugin.event({
  event: {
    type: 'file.edited',
    data: { path: '.env', content: 'SECRET=123', timestamp: Date.now() }
  }
});

Real-World Impact

Security: Prevent accidental sharing of credentials (env-protection plugin blocks .env file reads)

Productivity: Auto-notify on long-running commands (notify plugin sends system notifications)

Quality: Auto-format files on save (file.edited hook runs prettier)

Monitoring: Track tool usage patterns (tool.execute hooks log analytics)

Claude Code Event Mapping

When porting Claude Code hook behavior to OpenCode plugins, use these event mappings:

Claude HookOpenCode EventDescription
PreToolUsetool.execute.beforeRun before tool execution, can block
PostToolUsetool.execute.afterRun after tool execution
UserPromptSubmitmessage.* eventsProcess user prompts
SessionEndsession.idleSession completion

Example: Claude-like Hook Behavior

export const CompatiblePlugin = async (context) => {
  return {
    // Equivalent to Claude's PreToolUse hook
    'tool.execute.before': async (input, output) => {
      if (shouldBlock(input)) {
        throw new Error('Blocked by policy');
      }
    },

    // Equivalent to Claude's PostToolUse hook
    'tool.execute.after': async (result) => {
      console.log(`Tool completed: ${result.tool}`);
    },

    // Equivalent to Claude's SessionEnd hook
    event: async ({ event }) => {
      if (event.type === 'session.idle') {
        await cleanup();
      }
    }
  };
};

Plugin Composition

Combine multiple plugins using opencode-plugin-compose:

import { compose } from "opencode-plugin-compose";

const composedPlugin = compose([
  envProtectionPlugin,
  notifyPlugin,
  customToolsPlugin
]);
// Runs all hooks in sequence

Non-Convertibility Note

Important: OpenCode plugins cannot be directly converted from Claude Code hooks due to fundamental differences:

  • Event models differ: Claude has 4 hook events, OpenCode has 32+
  • Formats differ: Claude uses executable scripts, OpenCode uses JS/TS modules
  • Execution context differs: Different context objects and return value semantics

When porting Claude hooks to OpenCode plugins, you'll need to rewrite the logic using the OpenCode plugin API.


Schema Reference: packages/converters/schemas/opencode-plugin.schema.json

Documentation: https://opencode.ai/docs/plugins/

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.