feat: allow read tool to handle images (#3052)

packages/opencode/src/tool/read.ts

@@ -7,6 +7,8 @@ import { FileTime } from "../file/time"
 import DESCRIPTION from "./read.txt"
 import { Filesystem } from "../util/filesystem"
 import { Instance } from "../project/instance"
+import { Provider } from "../provider/provider"
+import { Identifier } from "../id/id"
 
 const DEFAULT_READ_LIMIT = 2000
 const MAX_LINE_LENGTH = 2000
@@ -23,6 +25,8 @@ export const ReadTool = Tool.define("read", {
     if (!path.isAbsolute(filepath)) {
       filepath = path.join(process.cwd(), filepath)
     }
+    const title = path.relative(Instance.worktree, filepath)
+
     if (!ctx.extra?.["bypassCwdCheck"] && !Filesystem.contains(Instance.directory, filepath)) {
       throw new Error(`File ${filepath} is not in the current working directory`)
     }
@@ -48,12 +52,45 @@ export const ReadTool = Tool.define("read", {
       throw new Error(`File not found: ${filepath}`)
     }
 
-    const limit = params.limit ?? DEFAULT_READ_LIMIT
-    const offset = params.offset || 0
     const isImage = isImageFile(filepath)
-    if (isImage) throw new Error(`This is an image file of type: ${isImage}\nUse a different tool to process images`)
+    const supportsImages = await (async () => {
+      if (!ctx.extra?.["providerID"] || !ctx.extra?.["modelID"]) return false
+      const providerID = ctx.extra["providerID"] as string
+      const modelID = ctx.extra["modelID"] as string
+      const model = await Provider.getModel(providerID, modelID).catch(() => undefined)
+      if (!model) return false
+      return model.info.modalities?.input?.includes("image") ?? false
+    })()
+    if (isImage) {
+      if (!supportsImages) {
+        throw new Error(`Failed to read image: ${filepath}, model may not be able to read images`)
+      }
+      const mime = file.type
+      const msg = "Image read successfully"
+      return {
+        title,
+        output: msg,
+        metadata: {
+          preview: msg,
+        },
+        attachments: [
+          {
+            id: Identifier.ascending("part"),
+            sessionID: ctx.sessionID,
+            messageID: ctx.messageID,
+            type: "file",
+            mime,
+            url: `data:${mime};base64,${Buffer.from(await file.bytes()).toString("base64")}`,
+          },
+        ],
+      }
+    }
+
     const isBinary = await isBinaryFile(filepath, file)
     if (isBinary) throw new Error(`Cannot read binary file: ${filepath}`)
+
+    const limit = params.limit ?? DEFAULT_READ_LIMIT
+    const offset = params.offset || 0
     const lines = await file.text().then((text) => text.split("\n"))
     const raw = lines.slice(offset, offset + limit).map((line) => {
       return line.length > MAX_LINE_LENGTH ? line.substring(0, MAX_LINE_LENGTH) + "..." : line
@@ -76,7 +113,7 @@ export const ReadTool = Tool.define("read", {
     FileTime.read(ctx.sessionID, filepath)
 
     return {
-      title: path.relative(Instance.worktree, filepath),
+      title,
       output,
       metadata: {
         preview,

packages/opencode/src/tool/read.txt

@@ -7,6 +7,6 @@ Usage:
 - You can optionally specify a line offset and limit (especially handy for long files), but it's recommended to read the whole file by not providing these parameters
 - Any lines longer than 2000 characters will be truncated
 - Results are returned using cat -n format, with line numbers starting at 1
-- This tool cannot read binary files, including images
-- You have the capability to call multiple tools in a single response. It is always better to speculatively read multiple files as a batch that are potentially useful. 
+- You have the capability to call multiple tools in a single response. It is always better to speculatively read multiple files as a batch that are potentially useful.
 - If you read a file that exists but has empty contents you will receive a system reminder warning in place of file contents.
+- You can read image files using this tool.

packages/opencode/src/tool/tool.ts

@@ -1,9 +1,11 @@
 import z from "zod/v4"
+import type { MessageV2 } from "../session/message-v2"
 
 export namespace Tool {
   interface Metadata {
     [key: string]: any
   }
+
   export type Context<M extends Metadata = Metadata> = {
     sessionID: string
     messageID: string
@@ -25,6 +27,7 @@ export namespace Tool {
         title: string
         metadata: M
         output: string
+        attachments?: MessageV2.FilePart[]
       }>
     }>
   }