opencode-minimax-easy-vision 1.2.2 → 1.3.0

package/README.md

@@ -6,6 +6,29 @@ Originally built for [MiniMax](https://www.minimax.io/) models, it can be config
 
 It restores the "paste and ask" workflow by automatically saving image assets and routing them through the [MiniMax Coding Plan MCP](https://github.com/MiniMax-AI/MiniMax-Coding-Plan-MCP)
 
+## Table of Contents
+
+* [Demo](#demo)
+* [The Problem](#the-problem)
+* [Prerequisites](#prerequisites)
+* [Installation](#installation)
+* [What This Plugin Does](#what-this-plugin-does)
+* [Supported Models](#supported-models)
+  * [Custom Model Configuration](#custom-model-configuration)
+    * [Locations (Priority Order)](#locations-priority-order)
+    * [Config Format](#config-format)
+    * [Pattern Syntax](#pattern-syntax)
+    * [Wildcard Rules](#wildcard-rules)
+    * [Configuration Examples](#configuration-examples)
+  * [Custom Image Analysis Tool](#custom-image-analysis-tool)
+  * [Custom Prompt Template](#custom-prompt-template)
+* [Supported Image Formats](#supported-image-formats)
+* [Usage](#usage)
+  * [Example Interaction](#example-interaction)
+* [Development](#development)
+* [License](#license)
+* [References](#references)
+
 ## Demo
 
 See how it works:
@@ -24,6 +47,36 @@ These models expect the MiniMax Coding Plan MCP's `understand_image` tool, which
 * **Manual steps**: You have to save screenshots manually, find the path, and reference it in your prompt.
 * **Broken flow**: The "paste and ask" experience available with Claude or GPT models is lost.
 
+## Prerequisites
+
+The MiniMax Coding Plan MCP server must be configured in your `opencode.json`:
+
+```json
+{
+  "mcp": {
+    "MiniMax": {
+      "type": "local",
+      "command": ["uvx", "minimax-coding-plan-mcp"],
+      "environment": {
+        "MINIMAX_API_KEY": "your-api-key-here",
+        "MINIMAX_API_HOST": "https://api.minimax.io"
+      }
+    }
+  }
+}
+```
+
+## Installation
+
+Add the plugin to the `plugin` array in your `opencode.json` file:
+
+```json
+{
+  "$schema": "https://opencode.ai/config.json",
+  "plugin": ["opencode-minimax-easy-vision"]
+}
+```
+
 ## What This Plugin Does
 
 This plugin automates the vision pipeline so you don't have to think about it.
@@ -39,14 +92,13 @@ This plugin automates the vision pipeline so you don't have to think about it.
 
 ## Supported Models
 
-By default, the plugin activates for MiniMax models:
+By default, the plugin activates for MiniMax models using the pattern `["minimax/*"]`:
 
-* **Provider ID** containing `minimax`
-* **Model ID** containing `minimax` or `abab`
+* `minimax/*` — all models from the `minimax` provider
 
 **Examples:**
+* `minimax/minimax-m2.5`
 * `minimax/minimax-m2.1`
-* `minimax/abab6.5s-chat`
 
 ### Custom Model Configuration
 
@@ -61,19 +113,18 @@ You can enable this for other models by creating a config file.
 
 ```json
 {
-  "models": ["minimax/*", "opencode/*", "*/glm-4.7-free"]
+  "models": ["minimax/*", "z-ai/*", "*/minimax-m2.5"]
 }
 ```
 
 #### Pattern Syntax
 
-| Pattern          | Matches                                 |
-| ---------------- | --------------------------------------- |
-| `*`              | Match ALL models                        |
-| `minimax/*`      | All models from the `minimax` provider  |
-| `*/glm-4.7-free` | Specific model from any provider        |
-| `opencode/*`     | All models from the `opencode` provider |
-| `*/abab*`        | Any model containing `abab`             |
+| Pattern               | Matches                                    |
+| --------------------- | ------------------------------------------ |
+| `*`                   | Match ALL models                           |
+| `minimax/*`           | All models from the `minimax` provider     |
+| `*/minimax-m2.5`      | Specific model from any provider           |
+| `z-ai/*`              | All models from the `z-ai` provider        |
 
 #### Wildcard Rules
 
@@ -82,7 +133,7 @@ You can enable this for other models by creating a config file.
 * `*` matches everything
 * `*text*` matches values containing `text`
 
-If the config is missing or empty, it defaults to MiniMax-only behavior.
+If the config is missing or empty, it defaults to `["minimax/*"]`.
 
 #### Configuration Examples
 
@@ -98,7 +149,7 @@ If the config is missing or empty, it defaults to MiniMax-only behavior.
 
 ```json
 {
-  "models": ["minimax/*", "opencode/*", "google/*"]
+  "models": ["minimax/*", "z-ai/*", "google/*"]
 }
 ```
 
@@ -106,84 +157,81 @@ If the config is missing or empty, it defaults to MiniMax-only behavior.
 
 ```json
 {
-  "models": ["minimax/*", "opencode/gpt-5-nano", "*/claude-3-7-sonnet*"]
+  "models": ["minimax/*", "z-ai/glm-4.7", "*/minimax-m2.5"]
 }
 ```
 
 ### Custom Image Analysis Tool
 
-By default, the plugin uses the `mcp_minimax_understand_image` tool from the MiniMax Coding Plan MCP. You can configure a different tool for image analysis:
-
-```json
-{
-  "models": ["*"],
-  "imageAnalysisTool": "mcp_openrouter_analyze_image"
-}
-```
+By default, the plugin uses the `mcp_minimax_understand_image` tool from the MiniMax Coding Plan MCP. You can configure a different tool for image analysis.
 
 > [!NOTE]
 > The `imageAnalysisTool` value is the **tool name**, not the MCP server name.
 >
 > MCP servers expose one or more tools—for example, the MiniMax Coding Plan MCP server exposes the
-> `mcp_minimax_understand_image` tool.
-
+> `mcp_minimax_understand_image` tool. OpenCode prefixes tool names with `mcp_<server-name>_`.
 
-This allows you to use tools from other MCP servers that provide image analysis capabilities, such as:
+#### Example: openrouter-image-mcp
 
-* [openrouter-image-mcp](https://github.com/JonathanJude/openrouter-image-mcp) - Uses OpenRouter with GPT-4V, Claude, Gemini
-* [mcp-image-recognition](https://github.com/mario-andreschak/mcp-image-recognition) - Uses Anthropic/OpenAI Vision APIs
-* [Peekaboo](https://github.com/steipete/Peekaboo) - macOS screenshot + AI analysis
+[openrouter-image-mcp](https://github.com/JonathanJude/openrouter-image-mcp) routes image analysis through OpenRouter, giving you access to any vision-capable model including free ones.
 
-The plugin will instruct the model to use the configured tool. The tool should accept an image file path as input.
+Add the MCP server to your `opencode.json`:
 
-## Supported Image Formats
+```json
+{
+  "mcp": {
+    "openrouter_image": {
+      "type": "local",
+      "command": ["npx", "openrouter-image-mcp"],
+      "environment": {
+        "OPENROUTER_API_KEY": "your-api-key-here",
+        "OPENROUTER_MODEL": "nvidia/nemotron-nano-12b-v2-vl:free"
+      }
+    }
+  }
+}
+```
 
-* PNG
-* JPEG
-* WebP
+Then configure the plugin to use it:
 
-*(Limited by the [MiniMax Coding Plan MCP](https://github.com/MiniMax-AI/MiniMax-Coding-Plan-MCP) `understand_image` tool.)*
+```json
+{
+  "models": ["minimax/*"],
+  "imageAnalysisTool": "mcp_openrouter_image_analyze_image"
+}
+```
 
-## Installation
+> [!TIP]
+> `nvidia/nemotron-nano-12b-v2-vl:free` is a free vision model on OpenRouter that requires no credits. 
 
-### Via npm
+### Custom Prompt Template
 
-Just add the plugin to the `plugin` array in your `opencode.json` file:
+By default, the plugin generates a fixed instruction prompt telling the model to use the image analysis tool. You can override this with a custom template:
 
 ```json
 {
-  "$schema": "https://opencode.ai/config.json",
-  "plugin": ["opencode-minimax-easy-vision"]
+  "promptTemplate": "I'm attaching {imageCount} image(s) for you to analyze.\n\nImages:\n{imageList}\n\nUse the `{toolName}` tool on each one.\n\nMy question: {userText}"
 }
 ```
 
-### From Local Source
+#### Available Variables
 
-1. Clone the repository.
-2. Build the plugin:
-   ```bash
-   npm install && npm run build
-   ```
-3. Copy the built `dist/index.js` into your OpenCode plugin directory.
+| Variable | Description |
+| -------------- | --------------------------------------------- |
+| `{imageList}` | Newline-separated list: `- Image 1: /path/to/file` |
+| `{imageCount}` | Number of images, e.g. `1`, `3` |
+| `{toolName}` | The configured MCP tool name |
+| `{userText}` | The user's original message text (may be empty) |
 
-## Prerequisites
+The template must contain at least one variable — if none are present, the plugin falls back to the default prompt.
 
-The MiniMax Coding Plan MCP server must be configured in your `opencode.json`:
+## Supported Image Formats
 
-```json
-{
-  "mcp": {
-    "MiniMax": {
-      "command": "uvx",
-      "args": ["minimax-coding-plan-mcp"],
-      "env": {
-        "MINIMAX_API_KEY": "your-api-key-here",
-        "MINIMAX_API_HOST": "https://api.minimax.io"
-      }
-    }
-  }
-}
-```
+* PNG
+* JPEG
+* WebP
+
+*(Limited by the [MiniMax Coding Plan MCP](https://github.com/MiniMax-AI/MiniMax-Coding-Plan-MCP) `understand_image` tool.)*
 
 ## Usage
 
@@ -202,12 +250,17 @@ The MiniMax Coding Plan MCP server must be configured in your `opencode.json`:
 
 ## Development
 
-```bash
-npm install
-npm run build
-```
-
-The built plugin will be available at `dist/index.js`
+1. Install dependencies and build:
+   ```bash
+   npm install
+   npm run build
+   ```
+2. The built plugin will be available at `dist/index.js`.
+3. Symlink it into the global plugin directory and restart OpenCode to pick up changes:
+   ```bash
+   mkdir -p ~/.config/opencode/plugin
+   ln -sf $(pwd)/dist/index.js ~/.config/opencode/plugin/minimax-easy-vision.js
+   ```
 
 ## License
 

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,qBAAqB,CAAC;AAqdlD,eAAO,MAAM,uBAAuB,EAAE,MA+DrC,CAAC;AAEF,eAAe,uBAAuB,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,qBAAqB,CAAC;AAqhBlD,eAAO,MAAM,uBAAuB,EAAE,MA+DrC,CAAC;AAEF,eAAe,uBAAuB,CAAC"}

package/dist/index.js

@@ -8,7 +8,13 @@ const PLUGIN_NAME = "minimax-easy-vision";
 const CONFIG_FILENAME = "opencode-minimax-easy-vision.json";
 const TEMP_DIR_NAME = "opencode-minimax-vision";
 const MAX_TOOL_NAME_LENGTH = 256;
-const DEFAULT_MODEL_PATTERNS = ["minimax/*", "*/abab*"];
+const PROMPT_TEMPLATE_VARIABLES = [
+    "{imageList}",
+    "{imageCount}",
+    "{toolName}",
+    "{userText}",
+];
+const DEFAULT_MODEL_PATTERNS = ["minimax/*"];
 const DEFAULT_IMAGE_ANALYSIS_TOOL = "mcp_minimax_understand_image";
 const SUPPORTED_MIME_TYPES = new Set([
     "image/png",
@@ -47,6 +53,16 @@ function parseImageAnalysisTool(value) {
         return undefined;
     return value;
 }
+function parsePromptTemplate(value) {
+    if (typeof value !== "string")
+        return undefined;
+    const trimmed = value.trim();
+    if (trimmed === "")
+        return undefined;
+    if (!PROMPT_TEMPLATE_VARIABLES.some((v) => trimmed.includes(v)))
+        return undefined;
+    return trimmed;
+}
 function parseConfigObject(raw) {
     if (!raw || typeof raw !== "object")
         return {};
@@ -54,6 +70,7 @@ function parseConfigObject(raw) {
     return {
         models: parseModelsArray(obj.models),
         imageAnalysisTool: parseImageAnalysisTool(obj.imageAnalysisTool),
+        promptTemplate: parsePromptTemplate(obj.promptTemplate),
     };
 }
 async function readConfigFile(configPath) {
@@ -84,7 +101,7 @@ async function loadPluginConfig(directory, log) {
     // Resolve models with precedence
     const modelsResult = selectWithPrecedence(projectConfig?.models, userConfig?.models, undefined);
     if (modelsResult.source !== "default") {
-        log(`Loaded models from ${modelsResult.source} config: ${modelsResult.value.join(", ")}`);
+        log(`Loaded models from ${modelsResult.source} config: ${modelsResult.value?.join(", ")}`);
     }
     else {
         log(`Using default models: ${DEFAULT_MODEL_PATTERNS.join(", ")}`);
@@ -97,9 +114,18 @@ async function loadPluginConfig(directory, log) {
     else {
         log(`Using default imageAnalysisTool: ${DEFAULT_IMAGE_ANALYSIS_TOOL}`);
     }
+    // Resolve promptTemplate with precedence
+    const templateResult = selectWithPrecedence(projectConfig?.promptTemplate, userConfig?.promptTemplate, undefined);
+    if (templateResult.source !== "default") {
+        log(`Using promptTemplate from ${templateResult.source} config (${templateResult.value?.length ?? 0} chars)`);
+    }
+    else {
+        log("Using default (hardcoded) injection prompt template");
+    }
     pluginConfig = {
         models: modelsResult.value,
         imageAnalysisTool: toolResult.value,
+        promptTemplate: templateResult.value,
     };
 }
 // Config: Accessors
@@ -109,6 +135,9 @@ function getConfiguredModels() {
 function getImageAnalysisTool() {
     return pluginConfig.imageAnalysisTool ?? DEFAULT_IMAGE_ANALYSIS_TOOL;
 }
+function getPromptTemplate() {
+    return pluginConfig.promptTemplate;
+}
 // Pattern Matching (supports wildcards: *, prefix*, *suffix, *contains*)
 function matchesWildcardPattern(pattern, value) {
     const p = pattern.toLowerCase();
@@ -204,7 +233,7 @@ async function handleDataUrl(url, filePart, log) {
         return { path: savedPath, mime: parsed.mime, partId: filePart.id };
     }
     catch (err) {
-        log(`Failed to save image: ${err}`);
+        log(`Failed to save image: ${err instanceof Error ? err.message : String(err)}`);
         return null;
     }
 }
@@ -246,7 +275,7 @@ async function processImagePart(filePart, log) {
     if (url.startsWith("http://") || url.startsWith("https://")) {
         return handleHttpUrl(url, filePart, log);
     }
-    log(`Unsupported URL scheme for part ${filePart.id}: ${url.substring(0, 50)}...`);
+    log(`Unsupported URL scheme for part ${filePart.id}: ${url.slice(0, 50)}...`);
     return null;
 }
 async function extractImagesFromParts(parts, log) {
@@ -267,13 +296,29 @@ async function extractImagesFromParts(parts, log) {
 // we replace them with text instructions that tell the model to use an
 // MCP tool (e.g., understand_image) with the file path or URL.
 // The user's original text is preserved as "User's request: ...".
+function applyPromptTemplate(template, vars) {
+    return template
+        .replace(/\{imageList\}/g, vars.imageList)
+        .replace(/\{imageCount\}/g, String(vars.imageCount))
+        .replace(/\{toolName\}/g, vars.toolName)
+        .replace(/\{userText\}/g, vars.userText);
+}
 function generateInjectionPrompt(images, userText, toolName) {
     if (images.length === 0)
         return userText;
-    const isSingle = images.length === 1;
     const imageList = images
         .map((img, idx) => `- Image ${idx + 1}: ${img.path}`)
         .join("\n");
+    const customTemplate = getPromptTemplate();
+    if (customTemplate !== undefined) {
+        return applyPromptTemplate(customTemplate, {
+            imageList,
+            imageCount: images.length,
+            toolName,
+            userText,
+        });
+    }
+    const isSingle = images.length === 1;
     const imageCountText = isSingle ? "an image" : `${images.length} images`;
     const imagePlural = isSingle ? "image is" : "images are";
     const analyzeText = isSingle ? "this image" : "each image";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "opencode-minimax-easy-vision",
-  "version": "1.2.2",
+  "version": "1.3.0",
   "description": "OpenCode plugin that enables vision support for Minimax models by saving pasted images and injecting MCP tool instructions",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -10,6 +10,8 @@
   ],
   "scripts": {
     "build": "tsc",
+    "format": "prettier --write src/**/*.ts",
+    "format:check": "prettier --check src/**/*.ts",
     "prepublishOnly": "npm run build"
   },
   "keywords": [
@@ -23,8 +25,9 @@
   "author": "devadathanmb",
   "license": "AGPL-3.0",
   "devDependencies": {
-    "@opencode-ai/plugin": "^1.1.25",
-    "@types/node": "^22.0.0",
+    "@opencode-ai/plugin": "^1.2.26",
+    "@types/node": "^25.0.0",
+    "prettier": "^3.8.1",
     "typescript": "^5.7.0"
   },
   "peerDependencies": {