diff --git a/loop/agent.go b/loop/agent.go
index 40bbb20..c98e7f7 100644
--- a/loop/agent.go
+++ b/loop/agent.go
@@ -1377,12 +1377,6 @@
 		claudetool.Think, claudetool.TodoRead, claudetool.TodoWrite, a.commitMessageStyleTool(), makeDoneTool(a.codereview),
 		a.codereview.Tool(), claudetool.AboutSketch,
 	}
-
-	// One-shot mode is non-interactive, multiple choice requires human response
-	if !a.config.OneShot {
-		convo.Tools = append(convo.Tools, multipleChoiceTool)
-	}
-
 	convo.Tools = append(convo.Tools, browserTools...)
 
 	// Add MCP tools if configured
@@ -1434,59 +1428,6 @@
 	return convo
 }
 
-var multipleChoiceTool = &llm.Tool{
-	Name:        "multiplechoice",
-	Description: "Present the user with an quick way to answer to your question using one of a short list of possible answers you would expect from the user.",
-	EndsTurn:    true,
-	InputSchema: json.RawMessage(`{
-  "type": "object",
-  "description": "The question and a list of answers you would expect the user to choose from.",
-  "properties": {
-    "question": {
-      "type": "string",
-      "description": "The text of the multiple-choice question you would like the user, e.g. 'What kinds of test cases would you like me to add?'"
-    },
-    "responseOptions": {
-      "type": "array",
-      "description": "The set of possible answers to let the user quickly choose from, e.g. ['Basic unit test coverage', 'Error return values', 'Malformed input'].",
-      "items": {
-        "type": "object",
-        "properties": {
-          "caption": {
-            "type": "string",
-            "description": "The caption, e.g. 'Basic coverage', 'Error return values', or 'Malformed input' for the response button. Do NOT include options for responses that would end the conversation like 'Ok', 'No thank you', 'This looks good'"
-          },
-          "responseText": {
-            "type": "string",
-            "description": "The full text of the response, e.g. 'Add unit tests for basic test coverage', 'Add unit tests for error return values', or 'Add unit tests for malformed input'"
-          }
-        },
-        "required": ["caption", "responseText"]
-      }
-    }
-  },
-  "required": ["question", "responseOptions"]
-}`),
-	Run: func(ctx context.Context, input json.RawMessage) llm.ToolOut {
-		// The Run logic for "multiplechoice" tool is a no-op on the server.
-		// The UI will present a list of options for the user to select from,
-		// and that's it as far as "executing" the tool_use goes.
-		// When the user *does* select one of the presented options, that
-		// responseText gets sent as a chat message on behalf of the user.
-		return llm.ToolOut{LLMContent: llm.TextContent("end your turn and wait for the user to respond")}
-	},
-}
-
-type MultipleChoiceOption struct {
-	Caption      string `json:"caption"`
-	ResponseText string `json:"responseText"`
-}
-
-type MultipleChoiceParams struct {
-	Question        string                 `json:"question"`
-	ResponseOptions []MultipleChoiceOption `json:"responseOptions"`
-}
-
 // branchExists reports whether branchName exists, either locally or in well-known remotes.
 func branchExists(dir, branchName string) bool {
 	refs := []string{
diff --git a/loop/agent_system_prompt.txt b/loop/agent_system_prompt.txt
index 2d60f72..0faf441 100644
--- a/loop/agent_system_prompt.txt
+++ b/loop/agent_system_prompt.txt
@@ -63,19 +63,6 @@
 Before modifying any file, you MUST proactively read and follow all guidance files in its directory and all parent directories.
 When guidance files conflict, more-deeply-nested files take precedence.
 Direct user instructions from the current conversation always take highest precedence.
-
-IMPORTANT: When the user provides feedback about how they want Sketch to behave, use the multiplechoice tool to ask whether to record this in a dear_llm.md file.
-If you think "I should remember this" - stop - and use multiplechoice instead.
-
-Always present all three of these options:
-
-+ 1. "Yes, for all future work" - Record in root dear_llm.md
-+ 2. "Yes, but only for directory X" - Record in X/dear_llm.md
-+ 3. "No" - Don't record this feedback
-
-When presenting this choice, the question must include a preview of exactly what would be written to the dear_llm.md file.
-For example: "Should I remember: 'Prefer table-driven tests over multiple separate test functions.'?"
-Changes to dear_llm.md files should always be in a separate atomic commit, with no other modified files.
 </customization>
 
 <guidance>
diff --git a/loop/testdata/agent_loop.httprr b/loop/testdata/agent_loop.httprr
index b2ac507..ee2cfa3 100644
--- a/loop/testdata/agent_loop.httprr
+++ b/loop/testdata/agent_loop.httprr
@@ -1,9 +1,9 @@
 httprr trace v1
-20094 2538
+18529 2497
 POST https://api.anthropic.com/v1/messages HTTP/1.1
 Host: api.anthropic.com
 User-Agent: Go-http-client/1.1
-Content-Length: 19896
+Content-Length: 18331
 Anthropic-Version: 2023-06-01
 Content-Type: application/json
 
@@ -291,45 +291,6 @@
    }
   },
   {
-   "name": "multiplechoice",
-   "description": "Present the user with an quick way to answer to your question using one of a short list of possible answers you would expect from the user.",
-   "input_schema": {
-    "type": "object",
-    "description": "The question and a list of answers you would expect the user to choose from.",
-    "properties": {
-     "question": {
-      "type": "string",
-      "description": "The text of the multiple-choice question you would like the user, e.g. 'What kinds of test cases would you like me to add?'"
-     },
-     "responseOptions": {
-      "type": "array",
-      "description": "The set of possible answers to let the user quickly choose from, e.g. ['Basic unit test coverage', 'Error return values', 'Malformed input'].",
-      "items": {
-       "type": "object",
-       "properties": {
-        "caption": {
-         "type": "string",
-         "description": "The caption, e.g. 'Basic coverage', 'Error return values', or 'Malformed input' for the response button. Do NOT include options for responses that would end the conversation like 'Ok', 'No thank you', 'This looks good'"
-        },
-        "responseText": {
-         "type": "string",
-         "description": "The full text of the response, e.g. 'Add unit tests for basic test coverage', 'Add unit tests for error return values', or 'Add unit tests for malformed input'"
-        }
-       },
-       "required": [
-        "caption",
-        "responseText"
-       ]
-      }
-     }
-    },
-    "required": [
-     "question",
-     "responseOptions"
-    ]
-   }
-  },
-  {
    "name": "browser_navigate",
    "description": "Navigate the browser to a specific URL and wait for page to load",
    "input_schema": {
@@ -587,24 +548,24 @@
 Anthropic-Organization-Id: 3c473a21-7208-450a-a9f8-80aebda45c1b
 Anthropic-Ratelimit-Input-Tokens-Limit: 2000000
 Anthropic-Ratelimit-Input-Tokens-Remaining: 2000000
-Anthropic-Ratelimit-Input-Tokens-Reset: 2025-07-23T00:51:23Z
+Anthropic-Ratelimit-Input-Tokens-Reset: 2025-07-23T18:04:13Z
 Anthropic-Ratelimit-Output-Tokens-Limit: 400000
 Anthropic-Ratelimit-Output-Tokens-Remaining: 400000
-Anthropic-Ratelimit-Output-Tokens-Reset: 2025-07-23T00:51:27Z
+Anthropic-Ratelimit-Output-Tokens-Reset: 2025-07-23T18:04:18Z
 Anthropic-Ratelimit-Requests-Limit: 4000
 Anthropic-Ratelimit-Requests-Remaining: 3999
-Anthropic-Ratelimit-Requests-Reset: 2025-07-23T00:51:20Z
+Anthropic-Ratelimit-Requests-Reset: 2025-07-23T18:04:12Z
 Anthropic-Ratelimit-Tokens-Limit: 2400000
 Anthropic-Ratelimit-Tokens-Remaining: 2400000
-Anthropic-Ratelimit-Tokens-Reset: 2025-07-23T00:51:23Z
+Anthropic-Ratelimit-Tokens-Reset: 2025-07-23T18:04:13Z
 Cf-Cache-Status: DYNAMIC
-Cf-Ray: 963730571c2eeb29-SJC
+Cf-Ray: 963d1951fa4567e9-SJC
 Content-Type: application/json
-Date: Wed, 23 Jul 2025 00:51:27 GMT
-Request-Id: req_011CRP2vuuf7xj5V5QqTaEuS
+Date: Wed, 23 Jul 2025 18:04:18 GMT
+Request-Id: req_011CRQPh7JZsCN9bRzBcZiu6
 Server: cloudflare
 Strict-Transport-Security: max-age=31536000; includeSubDomains; preload
 Via: 1.1 google
 X-Robots-Tag: none
 
-{"id":"msg_01MsAHgr98YWT6BpHDitnPPM","type":"message","role":"assistant","model":"claude-sonnet-4-20250514","content":[{"type":"text","text":"Here are the tools available to me:\n\n**File & Code Operations:**\n- `bash` - Execute shell commands\n- `patch` - Edit files with precise text modifications\n- `keyword_search` - Search codebase by keywords and concepts\n\n**Project Management:**\n- `todo_read` / `todo_write` - Manage task lists\n- `think` - Record thoughts and planning\n- `done` - Complete work with verification checklist\n\n**Git & Code Quality:**\n- `commit-message-style` - Get git commit message guidance\n- `codereview` - Run automated code review\n\n**Browser Automation:**\n- `browser_navigate` - Navigate to URLs\n- `browser_click` - Click elements\n- `browser_type` - Type into inputs\n- `browser_wait_for` - Wait for elements\n- `browser_get_text` - Read page text\n- `browser_eval` - Execute JavaScript\n- `browser_scroll_into_view` - Scroll to elements\n- `browser_resize` - Resize browser window\n- `browser_recent_console_logs` - Get console logs\n- `browser_clear_console_logs` - Clear console logs\n- `browser_take_screenshot` - Take screenshots\n\n**Utilities:**\n- `read_image` - Read and encode image files\n- `multiplechoice` - Present multiple choice questions\n- `about_sketch` - Get help with Sketch functionality"}],"stop_reason":"end_turn","stop_sequence":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4606,"cache_read_input_tokens":0,"output_tokens":323,"service_tier":"standard"}}
\ No newline at end of file
+{"id":"msg_01EGxSVYkNrBjKc5foxF4frd","type":"message","role":"assistant","model":"claude-sonnet-4-20250514","content":[{"type":"text","text":"Here are the tools available to me:\n\n**File & Code Management:**\n- `bash` - Execute shell commands\n- `patch` - Modify files with precise text edits\n- `keyword_search` - Search codebase by keywords\n\n**Task Management:**\n- `todo_read` - Read current todo list\n- `todo_write` - Create/manage task list\n- `think` - Take notes and form plans\n\n**Git & Code Quality:**\n- `commit-message-style` - Get git commit message guidance\n- `codereview` - Run automated code review\n- `done` - Mark completion with checklist verification\n\n**Browser Automation:**\n- `browser_navigate` - Navigate to URLs\n- `browser_click` - Click elements\n- `browser_type` - Type into inputs\n- `browser_wait_for` - Wait for elements\n- `browser_get_text` - Read element text\n- `browser_eval` - Execute JavaScript\n- `browser_scroll_into_view` - Scroll to elements\n- `browser_resize` - Resize browser window\n- `browser_take_screenshot` - Capture screenshots\n- `browser_recent_console_logs` - Get console logs\n- `browser_clear_console_logs` - Clear console logs\n\n**Utilities:**\n- `read_image` - Read and encode image files\n- `about_sketch` - Get help with Sketch functionality"}],"stop_reason":"end_turn","stop_sequence":null,"usage":{"input_tokens":3,"cache_creation_input_tokens":4261,"cache_read_input_tokens":0,"output_tokens":319,"service_tier":"standard"}}
\ No newline at end of file