release: v2.4.7

- Check if metadata is empty before using it
- Wrap output.output in structured object when metadata is missing
- Ensures plugin tools (call_omo_agent, background_task, task) that return strings are properly recorded in transcripts instead of empty {}

🤖 Generated with assistance of OhMyOpenCode

AGENTS.md

@@ -87,7 +87,7 @@ oh-my-opencode/
 | explore | opencode/grok-code | Fast codebase exploration, file patterns |
 | frontend-ui-ux-engineer | google/gemini-3-pro-preview | UI generation, design-focused |
 | document-writer | google/gemini-3-pro-preview | Technical documentation |
-| multimodal-looker | google/gemini-2.5-flash | PDF/image/diagram analysis |
+| multimodal-looker | google/gemini-3-flash | PDF/image/diagram analysis |
 
 ## COMMANDS
 

README.ja.md

@@ -317,12 +317,12 @@ opencode auth login
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```
 
-**利用可能なモデル名**: `google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-2.5-flash`, `google/gemini-2.5-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
+**利用可能なモデル名**: `google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-3-flash`, `google/gemini-3-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
 
 その後、認証を行います：
 
@@ -432,7 +432,7 @@ gh repo star code-yeongyu/oh-my-opencode
 - **explore** (`opencode/grok-code`): 高速なコードベース探索、ファイルパターンマッチング。Claude Code は Haiku を使用しますが、私たちは Grok を使います。現在無料であり、極めて高速で、ファイル探索タスクには十分な知能を備えているからです。Claude Code からインスピレーションを得ました。
 - **frontend-ui-ux-engineer** (`google/gemini-3-pro-preview`): 開発者に転身したデザイナーという設定です。素晴らしい UI を作ります。美しく独創的な UI コードを生成することに長けた Gemini を使用します。
 - **document-writer** (`google/gemini-3-pro-preview`): テクニカルライティングの専門家という設定です。Gemini は文筆家であり、流れるような文章を書きます。
-- **multimodal-looker** (`google/gemini-2.5-flash`): 視覚コンテンツ解釈のための専門エージェント。PDF、画像、図表を分析して情報を抽出します。
+- **multimodal-looker** (`google/gemini-3-flash`): 視覚コンテンツ解釈のための専門エージェント。PDF、画像、図表を分析して情報を抽出します。
 
 メインエージェントはこれらを自動的に呼び出しますが、明示的に呼び出すことも可能です：
 
@@ -675,7 +675,7 @@ Oh My OpenCode は以下の場所からフックを読み込んで実行しま
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```

README.ko.md

@@ -314,12 +314,12 @@ opencode auth login
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```
 
-**사용 가능한 모델 이름**: `google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-2.5-flash`, `google/gemini-2.5-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
+**사용 가능한 모델 이름**: `google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-3-flash`, `google/gemini-3-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
 
 그 후 인증:
 
@@ -429,7 +429,7 @@ gh repo star code-yeongyu/oh-my-opencode
 - **explore** (`opencode/grok-code`): 빠른 코드베이스 탐색, 파일 패턴 매칭. Claude Code는 Haiku를 쓰지만, 우리는 Grok을 씁니다. 현재 무료이고, 극도로 빠르며, 파일 탐색 작업에 충분한 지능을 갖췄기 때문입니다. Claude Code 에서 영감을 받았습니다.
 - **frontend-ui-ux-engineer** (`google/gemini-3-pro-preview`): 개발자로 전향한 디자이너라는 설정을 갖고 있습니다. 멋진 UI를 만듭니다. 아름답고 창의적인 UI 코드를 생성하는 데 탁월한 Gemini를 사용합니다.
 - **document-writer** (`google/gemini-3-pro-preview`): 기술 문서 전문가라는 설정을 갖고 있습니다. Gemini 는 문학가입니다. 글을 기가막히게 씁니다.
-- **multimodal-looker** (`google/gemini-2.5-flash`): 시각적 콘텐츠 해석을 위한 전문 에이전트. PDF, 이미지, 다이어그램을 분석하여 정보를 추출합니다.
+- **multimodal-looker** (`google/gemini-3-flash`): 시각적 콘텐츠 해석을 위한 전문 에이전트. PDF, 이미지, 다이어그램을 분석하여 정보를 추출합니다.
 
 각 에이전트는 메인 에이전트가 알아서 호출하지만, 명시적으로 요청할 수도 있습니다:
 
@@ -669,7 +669,7 @@ Schema 자동 완성이 지원됩니다:
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```

README.md

@@ -346,12 +346,12 @@ The `opencode-antigravity-auth` plugin uses different model names than the built
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```
 
-**Available model names**: `google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-2.5-flash`, `google/gemini-2.5-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
+**Available model names**: `google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-3-flash`, `google/gemini-3-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
 
 Then authenticate:
 
@@ -493,7 +493,7 @@ To remove oh-my-opencode:
 - **explore** (`opencode/grok-code`): Fast codebase exploration and pattern matching. Claude Code uses Haiku; we use Grok—it's free, blazing fast, and plenty smart for file traversal. Inspired by Claude Code.
 - **frontend-ui-ux-engineer** (`google/gemini-3-pro-preview`): A designer turned developer. Builds gorgeous UIs. Gemini excels at creative, beautiful UI code.
 - **document-writer** (`google/gemini-3-pro-preview`): Technical writing expert. Gemini is a wordsmith—writes prose that flows.
-- **multimodal-looker** (`google/gemini-2.5-flash`): Visual content specialist. Analyzes PDFs, images, diagrams to extract information.
+- **multimodal-looker** (`google/gemini-3-flash`): Visual content specialist. Analyzes PDFs, images, diagrams to extract information.
 
 The main agent invokes these automatically, but you can call them explicitly:
 
@@ -733,7 +733,7 @@ When using `opencode-antigravity-auth`, disable the built-in auth and override a
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```

README.zh-cn.md

@@ -325,12 +325,12 @@ opencode auth login
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```
 
-**可用模型名**：`google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-2.5-flash`, `google/gemini-2.5-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
+**可用模型名**：`google/gemini-3-pro-high`, `google/gemini-3-pro-medium`, `google/gemini-3-pro-low`, `google/gemini-3-flash`, `google/gemini-3-flash`, `google/gemini-3-flash-lite`, `google/claude-sonnet-4-5`, `google/claude-sonnet-4-5-thinking`, `google/claude-opus-4-5-thinking`, `google/gpt-oss-120b-medium`
 
 然后认证：
 
@@ -440,7 +440,7 @@ gh repo star code-yeongyu/oh-my-opencode
 - **explore** (`opencode/grok-code`)：极速代码库扫描、模式匹配。Claude Code 用 Haiku，我们用 Grok——免费、飞快、扫文件够用了。致敬 Claude Code。
 - **frontend-ui-ux-engineer** (`google/gemini-3-pro-preview`)：设计师出身的程序员。UI 做得那是真漂亮。Gemini 写这种创意美观的代码是一绝。
 - **document-writer** (`google/gemini-3-pro-preview`)：技术写作专家。Gemini 文笔好，写出来的东西读着顺畅。
-- **multimodal-looker** (`google/gemini-2.5-flash`)：视觉内容专家。PDF、图片、图表，看一眼就知道里头有啥。
+- **multimodal-looker** (`google/gemini-3-flash`)：视觉内容专家。PDF、图片、图表，看一眼就知道里头有啥。
 
 主 Agent 会自动调遣它们，你也可以亲自点名：
 
@@ -675,7 +675,7 @@ Agent 爽了，你自然也爽。但我还想直接让你爽。
   "agents": {
     "frontend-ui-ux-engineer": { "model": "google/gemini-3-pro-high" },
     "document-writer": { "model": "google/gemini-3-flash" },
-    "multimodal-looker": { "model": "google/gemini-2.5-flash" }
+    "multimodal-looker": { "model": "google/gemini-3-flash" }
   }
 }
 ```

package.json

@@ -1,6 +1,6 @@
 {
   "name": "oh-my-opencode",
-  "version": "2.4.4",
+  "version": "2.4.7",
   "description": "OpenCode plugin - custom agents (oracle, librarian) and enhanced features",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

src/agents/multimodal-looker.ts

@@ -4,7 +4,7 @@ export const multimodalLookerAgent: AgentConfig = {
   description:
     "Analyze media files (PDFs, images, diagrams) that require interpretation beyond raw text. Extracts specific information or summaries from documents, describes visual content. Use when you need analyzed/extracted data rather than literal file contents.",
   mode: "subagent",
-  model: "google/gemini-2.5-flash",
+  model: "google/gemini-3-flash",
   temperature: 0.1,
   tools: { write: false, edit: false, bash: false, background_task: false },
   prompt: `You interpret media files that cannot be read as plain text.

src/hooks/claude-code-hooks/index.ts

@@ -184,7 +184,13 @@ export function createClaudeCodeHooksHook(ctx: PluginInput, config: PluginConfig
 
       const cachedInput = getToolInput(input.sessionID, input.tool, input.callID) || {}
 
-      recordToolResult(input.sessionID, input.tool, cachedInput, (output.metadata as Record<string, unknown>) || {})
+      // Use metadata if available and non-empty, otherwise wrap output.output in a structured object
+      // This ensures plugin tools (call_omo_agent, background_task, task) that return strings
+      // get their results properly recorded in transcripts instead of empty {}
+      const metadata = output.metadata as Record<string, unknown> | undefined
+      const hasMetadata = metadata && typeof metadata === "object" && Object.keys(metadata).length > 0
+      const toolOutput = hasMetadata ? metadata : { output: output.output }
+      recordToolResult(input.sessionID, input.tool, cachedInput, toolOutput)
 
       if (!isHookDisabled(config, "PostToolUse")) {
         const postClient: PostToolUseClient = {

src/hooks/keyword-detector/index.ts

@@ -25,11 +25,6 @@ export function createKeywordDetectorHook() {
       const isFirstMessage = !sessionFirstMessageProcessed.has(input.sessionID)
       sessionFirstMessageProcessed.add(input.sessionID)
 
-      if (isFirstMessage) {
-        log("Skipping keyword detection on first message for title generation", { sessionID: input.sessionID })
-        return
-      }
-
       const promptText = extractPromptText(output.parts)
       const messages = detectKeywords(promptText)
 
@@ -37,6 +32,19 @@ export function createKeywordDetectorHook() {
         return
       }
 
+      const context = messages.join("\n")
+
+      // First message: transform parts directly (for title generation compatibility)
+      if (isFirstMessage) {
+        log(`Keywords detected on first message, transforming parts directly`, { sessionID: input.sessionID, keywordCount: messages.length })
+        const idx = output.parts.findIndex((p) => p.type === "text" && p.text)
+        if (idx >= 0) {
+          output.parts[idx].text = `${context}\n\n---\n\n${output.parts[idx].text ?? ""}`
+        }
+        return
+      }
+
+      // Subsequent messages: inject as separate message
       log(`Keywords detected: ${messages.length}`, { sessionID: input.sessionID })
 
       const message = output.message as {
@@ -46,7 +54,6 @@ export function createKeywordDetectorHook() {
         tools?: Record<string, boolean>
       }
 
-      const context = messages.join("\n")
       log(`[keyword-detector] Injecting context for ${messages.length} keywords`, { sessionID: input.sessionID, contextLength: context.length })
       const success = injectHookMessage(input.sessionID, context, {
         agent: message.agent,

src/hooks/non-interactive-env/constants.ts

@@ -14,4 +14,56 @@ export const NON_INTERACTIVE_ENV: Record<string, string> = {
   // Block pagers
   GIT_PAGER: "cat",
   PAGER: "cat",
+  // NPM non-interactive
+  npm_config_yes: "true",
+  // Pip non-interactive
+  PIP_NO_INPUT: "1",
+  // Yarn non-interactive
+  YARN_ENABLE_IMMUTABLE_INSTALLS: "false",
 }
+
+/**
+ * Shell command guidance for non-interactive environments.
+ * These patterns should be followed to avoid hanging on user input.
+ */
+export const SHELL_COMMAND_PATTERNS = {
+  // Package managers - always use non-interactive flags
+  npm: {
+    bad: ["npm init", "npm install (prompts)"],
+    good: ["npm init -y", "npm install --yes"],
+  },
+  apt: {
+    bad: ["apt-get install pkg"],
+    good: ["apt-get install -y pkg", "DEBIAN_FRONTEND=noninteractive apt-get install pkg"],
+  },
+  pip: {
+    bad: ["pip install pkg (with prompts)"],
+    good: ["pip install --no-input pkg", "PIP_NO_INPUT=1 pip install pkg"],
+  },
+  // Git operations - always provide messages/flags
+  git: {
+    bad: ["git commit", "git merge branch", "git add -p", "git rebase -i"],
+    good: ["git commit -m 'msg'", "git merge --no-edit branch", "git add .", "git rebase --no-edit"],
+  },
+  // System commands - force flags
+  system: {
+    bad: ["rm file (prompts)", "cp a b (prompts)", "ssh host"],
+    good: ["rm -f file", "cp -f a b", "ssh -o BatchMode=yes host", "unzip -o file.zip"],
+  },
+  // Banned commands - will always hang
+  banned: [
+    "vim", "nano", "vi", "emacs",           // Editors
+    "less", "more", "man",                   // Pagers
+    "python (REPL)", "node (REPL)",          // REPLs without -c/-e
+    "git add -p", "git rebase -i",           // Interactive git modes
+  ],
+  // Workarounds for scripts that require input
+  workarounds: {
+    yesPipe: "yes | ./script.sh",
+    heredoc: `./script.sh <<EOF
+option1
+option2
+EOF`,
+    expectAlternative: "Use environment variables or config files instead of expect",
+  },
+} as const

src/hooks/non-interactive-env/index.ts

@@ -1,15 +1,28 @@
 import type { PluginInput } from "@opencode-ai/plugin"
-import { HOOK_NAME, NON_INTERACTIVE_ENV } from "./constants"
+import { HOOK_NAME, NON_INTERACTIVE_ENV, SHELL_COMMAND_PATTERNS } from "./constants"
 import { log } from "../../shared"
 
 export * from "./constants"
 export * from "./types"
 
+const BANNED_COMMAND_PATTERNS = SHELL_COMMAND_PATTERNS.banned
+  .filter((cmd) => !cmd.includes("("))
+  .map((cmd) => new RegExp(`\\b${cmd}\\b`))
+
+function detectBannedCommand(command: string): string | undefined {
+  for (let i = 0; i < BANNED_COMMAND_PATTERNS.length; i++) {
+    if (BANNED_COMMAND_PATTERNS[i].test(command)) {
+      return SHELL_COMMAND_PATTERNS.banned[i]
+    }
+  }
+  return undefined
+}
+
 export function createNonInteractiveEnvHook(_ctx: PluginInput) {
   return {
     "tool.execute.before": async (
       input: { tool: string; sessionID: string; callID: string },
-      output: { args: Record<string, unknown> }
+      output: { args: Record<string, unknown>; message?: string }
     ): Promise<void> => {
       if (input.tool.toLowerCase() !== "bash") {
         return
@@ -25,6 +38,11 @@ export function createNonInteractiveEnvHook(_ctx: PluginInput) {
         ...NON_INTERACTIVE_ENV,
       }
 
+      const bannedCmd = detectBannedCommand(command)
+      if (bannedCmd) {
+        output.message = `⚠️ Warning: '${bannedCmd}' is an interactive command that may hang in non-interactive environments.`
+      }
+
       log(`[${HOOK_NAME}] Set non-interactive environment variables`, {
         sessionID: input.sessionID,
         env: NON_INTERACTIVE_ENV,

src/hooks/preemptive-compaction/index.ts

@@ -1,3 +1,5 @@
+import { existsSync, readdirSync } from "node:fs"
+import { join } from "node:path"
 import type { PluginInput } from "@opencode-ai/plugin"
 import type { ExperimentalConfig } from "../../config"
 import type { PreemptiveCompactionState, TokenInfo } from "./types"
@@ -6,6 +8,10 @@ import {
   MIN_TOKENS_FOR_COMPACTION,
   COMPACTION_COOLDOWN_MS,
 } from "./constants"
+import {
+  findNearestMessageWithFields,
+  MESSAGE_STORAGE,
+} from "../../features/hook-message-injector"
 import { log } from "../../shared/logger"
 
 export interface SummarizeContext {
@@ -48,6 +54,20 @@ function isSupportedModel(modelID: string): boolean {
   return CLAUDE_MODEL_PATTERN.test(modelID)
 }
 
+function getMessageDir(sessionID: string): string | null {
+  if (!existsSync(MESSAGE_STORAGE)) return null
+
+  const directPath = join(MESSAGE_STORAGE, sessionID)
+  if (existsSync(directPath)) return directPath
+
+  for (const dir of readdirSync(MESSAGE_STORAGE)) {
+    const sessionPath = join(MESSAGE_STORAGE, dir, sessionID)
+    if (existsSync(sessionPath)) return sessionPath
+  }
+
+  return null
+}
+
 function createState(): PreemptiveCompactionState {
   return {
     lastCompactionTime: new Map(),
@@ -164,9 +184,15 @@ export function createPreemptiveCompactionHook(
 
       setTimeout(async () => {
         try {
+          const messageDir = getMessageDir(sessionID)
+          const storedMessage = messageDir ? findNearestMessageWithFields(messageDir) : null
+
           await ctx.client.session.promptAsync({
             path: { id: sessionID },
-            body: { parts: [{ type: "text", text: "Continue" }] },
+            body: {
+              agent: storedMessage?.agent,
+              parts: [{ type: "text", text: "Continue" }],
+            },
             query: { directory: ctx.directory },
           })
         } catch {}
@@ -222,6 +248,21 @@ export function createPreemptiveCompactionHook(
         if (assistants.length === 0) return
 
         const lastAssistant = assistants[assistants.length - 1]
+
+        if (!lastAssistant.providerID || !lastAssistant.modelID) {
+          const messageDir = getMessageDir(sessionID)
+          const storedMessage = messageDir ? findNearestMessageWithFields(messageDir) : null
+          if (storedMessage?.model?.providerID && storedMessage?.model?.modelID) {
+            lastAssistant.providerID = storedMessage.model.providerID
+            lastAssistant.modelID = storedMessage.model.modelID
+            log("[preemptive-compaction] using stored message model info", {
+              sessionID,
+              providerID: lastAssistant.providerID,
+              modelID: lastAssistant.modelID,
+            })
+          }
+        }
+
         await checkAndTriggerCompaction(sessionID, lastAssistant)
       } catch {}
     }

src/tools/index.ts

@@ -20,7 +20,6 @@ import {
 import { grep } from "./grep"
 import { glob } from "./glob"
 import { slashcommand } from "./slashcommand"
-import { skill } from "./skill"
 
 export { interactive_bash, startBackgroundCheck as startTmuxCheck } from "./interactive-bash"
 export { getTmuxPath } from "./interactive-bash/utils"
@@ -64,5 +63,4 @@ export const builtinTools = {
   grep,
   glob,
   slashcommand,
-  skill,
 }

src/tools/look-at/tools.ts

@@ -1,8 +1,33 @@
+import { extname, basename } from "node:path"
 import { tool, type PluginInput } from "@opencode-ai/plugin"
 import { LOOK_AT_DESCRIPTION, MULTIMODAL_LOOKER_AGENT } from "./constants"
 import type { LookAtArgs } from "./types"
 import { log } from "../../shared/logger"
 
+function inferMimeType(filePath: string): string {
+  const ext = extname(filePath).toLowerCase()
+  const mimeTypes: Record<string, string> = {
+    ".jpg": "image/jpeg",
+    ".jpeg": "image/jpeg",
+    ".png": "image/png",
+    ".gif": "image/gif",
+    ".webp": "image/webp",
+    ".svg": "image/svg+xml",
+    ".bmp": "image/bmp",
+    ".ico": "image/x-icon",
+    ".pdf": "application/pdf",
+    ".txt": "text/plain",
+    ".md": "text/markdown",
+    ".json": "application/json",
+    ".xml": "application/xml",
+    ".html": "text/html",
+    ".css": "text/css",
+    ".js": "text/javascript",
+    ".ts": "text/typescript",
+  }
+  return mimeTypes[ext] || "application/octet-stream"
+}
+
 export function createLookAt(ctx: PluginInput) {
   return tool({
     description: LOOK_AT_DESCRIPTION,
@@ -13,12 +38,14 @@ export function createLookAt(ctx: PluginInput) {
     async execute(args: LookAtArgs, toolContext) {
       log(`[look_at] Analyzing file: ${args.file_path}, goal: ${args.goal}`)
 
+      const mimeType = inferMimeType(args.file_path)
+      const filename = basename(args.file_path)
+
       const prompt = `Analyze this file and extract the requested information.
 
-File path: ${args.file_path}
 Goal: ${args.goal}
 
-Read the file using the Read tool, then provide ONLY the extracted information that matches the goal.
+Provide ONLY the extracted information that matches the goal.
 Be thorough on what was requested, concise on everything else.
 If the requested information is not found, clearly state what is missing.`
 
@@ -38,7 +65,7 @@ If the requested information is not found, clearly state what is missing.`
       const sessionID = createResult.data.id
       log(`[look_at] Created session: ${sessionID}`)
 
-      log(`[look_at] Sending prompt to session ${sessionID}`)
+      log(`[look_at] Sending prompt with file passthrough to session ${sessionID}`)
       await ctx.client.session.prompt({
         path: { id: sessionID },
         body: {
@@ -47,8 +74,12 @@ If the requested information is not found, clearly state what is missing.`
             task: false,
             call_omo_agent: false,
             look_at: false,
+            read: false,
           },
-          parts: [{ type: "text", text: prompt }],
+          parts: [
+            { type: "text", text: prompt },
+            { type: "file", mime: mimeType, url: `file://${args.file_path}`, filename },
+          ],
         },
       })
 

src/tools/skill/index.ts

@@ -1,2 +0,0 @@
-export * from "./types"
-export { skill } from "./tools"

src/tools/skill/tools.ts

@@ -1,304 +0,0 @@
-import { tool } from "@opencode-ai/plugin"
-import { existsSync, readdirSync, readFileSync } from "fs"
-import { homedir } from "os"
-import { join, basename } from "path"
-import { z } from "zod/v4"
-import { parseFrontmatter, resolveCommandsInText } from "../../shared"
-import { resolveSymlink } from "../../shared/file-utils"
-import { SkillFrontmatterSchema } from "./types"
-import type { SkillScope, SkillMetadata, SkillInfo, LoadedSkill, SkillFrontmatter } from "./types"
-
-function parseSkillFrontmatter(data: Record<string, unknown>): SkillFrontmatter {
-  return {
-    name: typeof data.name === "string" ? data.name : "",
-    description: typeof data.description === "string" ? data.description : "",
-    license: typeof data.license === "string" ? data.license : undefined,
-    "allowed-tools": Array.isArray(data["allowed-tools"]) ? data["allowed-tools"] : undefined,
-    metadata:
-      typeof data.metadata === "object" && data.metadata !== null
-        ? (data.metadata as Record<string, string>)
-        : undefined,
-  }
-}
-
-function discoverSkillsFromDir(
-  skillsDir: string,
-  scope: SkillScope
-): Array<{ name: string; description: string; scope: SkillScope }> {
-  if (!existsSync(skillsDir)) {
-    return []
-  }
-
-  const entries = readdirSync(skillsDir, { withFileTypes: true })
-  const skills: Array<{ name: string; description: string; scope: SkillScope }> = []
-
-  for (const entry of entries) {
-    if (entry.name.startsWith(".")) continue
-
-    const skillPath = join(skillsDir, entry.name)
-
-    if (entry.isDirectory() || entry.isSymbolicLink()) {
-      const resolvedPath = resolveSymlink(skillPath)
-
-      const skillMdPath = join(resolvedPath, "SKILL.md")
-      if (!existsSync(skillMdPath)) continue
-
-      try {
-        const content = readFileSync(skillMdPath, "utf-8")
-        const { data } = parseFrontmatter(content)
-
-        skills.push({
-          name: data.name || entry.name,
-          description: data.description || "",
-          scope,
-        })
-      } catch {
-        continue
-      }
-    }
-  }
-
-  return skills
-}
-
-function discoverSkillsSync(): Array<{ name: string; description: string; scope: SkillScope }> {
-  const userSkillsDir = join(homedir(), ".claude", "skills")
-  const projectSkillsDir = join(process.cwd(), ".claude", "skills")
-
-  const userSkills = discoverSkillsFromDir(userSkillsDir, "user")
-  const projectSkills = discoverSkillsFromDir(projectSkillsDir, "project")
-
-  return [...projectSkills, ...userSkills]
-}
-
-const availableSkills = discoverSkillsSync()
-const skillListForDescription = availableSkills
-  .map((s) => `- ${s.name}: ${s.description} (${s.scope})`)
-  .join("\n")
-
-async function parseSkillMd(skillPath: string): Promise<SkillInfo | null> {
-  const resolvedPath = resolveSymlink(skillPath)
-  const skillMdPath = join(resolvedPath, "SKILL.md")
-
-  if (!existsSync(skillMdPath)) {
-    return null
-  }
-
-  try {
-    let content = readFileSync(skillMdPath, "utf-8")
-    content = await resolveCommandsInText(content)
-    const { data, body } = parseFrontmatter(content)
-
-    const frontmatter = parseSkillFrontmatter(data)
-
-    const metadata: SkillMetadata = {
-      name: frontmatter.name || basename(skillPath),
-      description: frontmatter.description,
-      license: frontmatter.license,
-      allowedTools: frontmatter["allowed-tools"],
-      metadata: frontmatter.metadata,
-    }
-
-    const referencesDir = join(resolvedPath, "references")
-    const scriptsDir = join(resolvedPath, "scripts")
-    const assetsDir = join(resolvedPath, "assets")
-
-    const references = existsSync(referencesDir)
-      ? readdirSync(referencesDir).filter((f) => !f.startsWith("."))
-      : []
-
-    const scripts = existsSync(scriptsDir)
-      ? readdirSync(scriptsDir).filter((f) => !f.startsWith(".") && !f.startsWith("__"))
-      : []
-
-    const assets = existsSync(assetsDir)
-      ? readdirSync(assetsDir).filter((f) => !f.startsWith("."))
-      : []
-
-    return {
-      name: metadata.name,
-      path: resolvedPath,
-      basePath: resolvedPath,
-      metadata,
-      content: body,
-      references,
-      scripts,
-      assets,
-    }
-  } catch {
-    return null
-  }
-}
-
-async function discoverSkillsFromDirAsync(skillsDir: string): Promise<SkillInfo[]> {
-  if (!existsSync(skillsDir)) {
-    return []
-  }
-
-  const entries = readdirSync(skillsDir, { withFileTypes: true })
-  const skills: SkillInfo[] = []
-
-  for (const entry of entries) {
-    if (entry.name.startsWith(".")) continue
-
-    const skillPath = join(skillsDir, entry.name)
-
-    if (entry.isDirectory() || entry.isSymbolicLink()) {
-      const skillInfo = await parseSkillMd(skillPath)
-      if (skillInfo) {
-        skills.push(skillInfo)
-      }
-    }
-  }
-
-  return skills
-}
-
-async function discoverSkills(): Promise<SkillInfo[]> {
-  const userSkillsDir = join(homedir(), ".claude", "skills")
-  const projectSkillsDir = join(process.cwd(), ".claude", "skills")
-
-  const userSkills = await discoverSkillsFromDirAsync(userSkillsDir)
-  const projectSkills = await discoverSkillsFromDirAsync(projectSkillsDir)
-
-  return [...projectSkills, ...userSkills]
-}
-
-function findMatchingSkills(skills: SkillInfo[], query: string): SkillInfo[] {
-  const queryLower = query.toLowerCase()
-  const queryTerms = queryLower.split(/\s+/).filter(Boolean)
-
-  return skills
-    .map((skill) => {
-      let score = 0
-      const nameLower = skill.metadata.name.toLowerCase()
-      const descLower = skill.metadata.description.toLowerCase()
-
-      if (nameLower === queryLower) score += 100
-      if (nameLower.includes(queryLower)) score += 50
-
-      for (const term of queryTerms) {
-        if (nameLower.includes(term)) score += 20
-        if (descLower.includes(term)) score += 10
-      }
-
-      return { skill, score }
-    })
-    .filter(({ score }) => score > 0)
-    .sort((a, b) => b.score - a.score)
-    .map(({ skill }) => skill)
-}
-
-async function loadSkillWithReferences(
-  skill: SkillInfo,
-  includeRefs: boolean
-): Promise<LoadedSkill> {
-  const referencesLoaded: Array<{ path: string; content: string }> = []
-
-  if (includeRefs && skill.references.length > 0) {
-    for (const ref of skill.references) {
-      const refPath = join(skill.path, "references", ref)
-      try {
-        let content = readFileSync(refPath, "utf-8")
-        content = await resolveCommandsInText(content)
-        referencesLoaded.push({ path: ref, content })
-      } catch {
-        // Skip unreadable references
-      }
-    }
-  }
-
-  return {
-    name: skill.name,
-    metadata: skill.metadata,
-    basePath: skill.basePath,
-    body: skill.content,
-    referencesLoaded,
-  }
-}
-
-function formatSkillList(skills: SkillInfo[]): string {
-  if (skills.length === 0) {
-    return "No skills found in ~/.claude/skills/"
-  }
-
-  const lines = ["# Available Skills\n"]
-
-  for (const skill of skills) {
-    lines.push(`- **${skill.metadata.name}**: ${skill.metadata.description || "(no description)"}`)
-  }
-
-  lines.push(`\n**Total**: ${skills.length} skills`)
-  return lines.join("\n")
-}
-
-function formatLoadedSkills(loadedSkills: LoadedSkill[]): string {
-  if (loadedSkills.length === 0) {
-    return "No skills loaded."
-  }
-
-  const skill = loadedSkills[0]
-  const sections: string[] = []
-
-  sections.push(`Base directory for this skill: ${skill.basePath}/`)
-  sections.push("")
-  sections.push(skill.body.trim())
-
-  if (skill.referencesLoaded.length > 0) {
-    sections.push("\n---\n### Loaded References\n")
-    for (const ref of skill.referencesLoaded) {
-      sections.push(`#### ${ref.path}\n`)
-      sections.push("```")
-      sections.push(ref.content.trim())
-      sections.push("```\n")
-    }
-  }
-
-  sections.push(`\n---\n**Launched skill**: ${skill.metadata.name}`)
-
-  return sections.join("\n")
-}
-
-export const skill = tool({
-  description: `Execute a skill within the main conversation.
-
-When you invoke a skill, the skill's prompt will expand and provide detailed instructions on how to complete the task.
-
-Available Skills:
-${skillListForDescription}`,
-
-  args: {
-    skill: tool.schema
-      .string()
-      .describe(
-        "The skill name or search query to find and load. Can be exact skill name (e.g., 'python-programmer') or keywords (e.g., 'python', 'plan')."
-      ),
-  },
-
-  async execute(args) {
-    const skills = await discoverSkills()
-
-    if (!args.skill) {
-      return formatSkillList(skills) + "\n\nProvide a skill name to load."
-    }
-
-    const matchingSkills = findMatchingSkills(skills, args.skill)
-
-    if (matchingSkills.length === 0) {
-      return (
-        `No skills found matching "${args.skill}".\n\n` +
-        formatSkillList(skills) +
-        "\n\nTry a different skill name."
-      )
-    }
-
-    const loadedSkills: LoadedSkill[] = []
-
-    for (const skillInfo of matchingSkills.slice(0, 3)) {
-      const loaded = await loadSkillWithReferences(skillInfo, true)
-      loadedSkills.push(loaded)
-    }
-
-    return formatLoadedSkills(loadedSkills)
-  },
-})

src/tools/skill/types.ts

@@ -1,47 +0,0 @@
-import { z } from "zod/v4"
-
-export type SkillScope = "user" | "project"
-
-/**
- * Zod schema for skill frontmatter validation
- * Following Anthropic Agent Skills Specification v1.0
- */
-export const SkillFrontmatterSchema = z.object({
-  name: z
-    .string()
-    .regex(/^[a-z0-9-]+$/, "Name must be lowercase alphanumeric with hyphens only")
-    .min(1, "Name cannot be empty"),
-  description: z.string().min(20, "Description must be at least 20 characters for discoverability"),
-  license: z.string().optional(),
-  "allowed-tools": z.array(z.string()).optional(),
-  metadata: z.record(z.string(), z.string()).optional(),
-})
-
-export type SkillFrontmatter = z.infer<typeof SkillFrontmatterSchema>
-
-export interface SkillMetadata {
-  name: string
-  description: string
-  license?: string
-  allowedTools?: string[]
-  metadata?: Record<string, string>
-}
-
-export interface SkillInfo {
-  name: string
-  path: string
-  basePath: string
-  metadata: SkillMetadata
-  content: string
-  references: string[]
-  scripts: string[]
-  assets: string[]
-}
-
-export interface LoadedSkill {
-  name: string
-  metadata: SkillMetadata
-  basePath: string
-  body: string
-  referencesLoaded: Array<{ path: string; content: string }>
-}