From a668860b8690899b8ada141bc7f87be0f7fb9f87 Mon Sep 17 00:00:00 2001
From: YeonGyu-Kim <code.yeon.gyu@gmail.com>
Date: Wed, 11 Mar 2026 21:56:19 +0900
Subject: [PATCH] fix: adjust vision capability check to not block when no
 model resolved

- Only block when a resolved model is explicitly not vision-capable
- Set up vision cache in model passthrough test for proper isolation
---
 src/tools/look-at/tools.test.ts | 3 +++
 src/tools/look-at/tools.ts      | 8 ++++----
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/src/tools/look-at/tools.test.ts b/src/tools/look-at/tools.test.ts
index 3312fbbd2..45a35f9a3 100644
--- a/src/tools/look-at/tools.test.ts
+++ b/src/tools/look-at/tools.test.ts
@@ -1,5 +1,6 @@
 import { describe, expect, test, mock } from "bun:test"
 import type { ToolContext } from "@opencode-ai/plugin/tool"
+import { setVisionCapableModelsCache } from "../../shared/vision-capable-models-cache"
 import { normalizeArgs, validateArgs, createLookAt } from "./tools"
 
 describe("look-at tool", () => {
@@ -255,6 +256,8 @@ describe("look-at tool", () => {
     // when LookAt tool executed
     // then model info should be passed to sync prompt
     test("passes multimodal-looker model to sync prompt when available", async () => {
+      setVisionCapableModelsCache(new Map([["google/gemini-3-flash", { providerID: "google", modelID: "gemini-3-flash" }]]))
+
       let promptBody: any
 
       const mockClient = {
diff --git a/src/tools/look-at/tools.ts b/src/tools/look-at/tools.ts
index 3fe5e4e11..a5fdb6075 100644
--- a/src/tools/look-at/tools.ts
+++ b/src/tools/look-at/tools.ts
@@ -148,11 +148,11 @@ Be thorough on what was requested, concise on everything else.
 If the requested information is not found, clearly state what is missing.`
 
       const { agentModel, agentVariant } = await resolveMultimodalLookerAgentMetadata(ctx)
-      if (!agentModel || !isVisionCapableResolvedModel(agentModel)) {
-        log("[look_at] No vision-capable multimodal-looker model resolved", {
+      if (agentModel && !isVisionCapableResolvedModel(agentModel)) {
+        log("[look_at] Resolved model is not vision-capable, blocking", {
           resolvedModel: agentModel,
         })
-        return "Error: No vision-capable multimodal-looker model available"
+        return "Error: Resolved multimodal-looker model is not vision-capable"
       }
 
       log(`[look_at] Creating session with parent: ${toolContext.sessionID}`)
@@ -204,7 +204,7 @@ Original error: ${createResult.error}`
               { type: "text", text: prompt },
               filePart,
             ],
-            model: { providerID: agentModel.providerID, modelID: agentModel.modelID },
+            ...(agentModel ? { model: { providerID: agentModel.providerID, modelID: agentModel.modelID } } : {}),
             ...(agentVariant ? { variant: agentVariant } : {}),
           },
         })