fix: 修复 anthropic 煞笔的四个 bug (#352)

* fix: 移除文件编辑前必须先读取的限制

移除 FileEditTool 和 FileWriteTool 中的 "read before edit" 校验，
允许直接编辑未读取过的文件。保留文件修改过期检测。

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* docs: 更新 teach-me 自动写 note 笔记的功能

* fix: 修复 DeepSeek V4 reasoning_content 回传导致的 400 错误

- 扩大模型名称检测范围，匹配所有 deepseek 模型（V4、R1 等）
- 始终保留 thinking blocks 为 reasoning_content 回传给 API
- 移除有 bug 的 turn boundary 剥离逻辑

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix: Opus 4.6/4.7 默认推理 effort 从 medium 改为 high

Pro 和 Max/Team 订阅者的 Opus 默认 effort 之前被降级为 medium，
导致用户感知模型「变笨」。恢复为 high。

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix: 移除 thinkingClearLatched sticky-on 机制

空闲超过 1 小时后 thinkingClearLatched 会被触发且永不重置，
导致每轮 API 调用都清除 thinking 历史。完整移除该 latch 机制，
clearAllThinking 硬编码为 false。

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix: 移除 numeric_length_anchors 系统指令

删除「工具调用间文字 ≤25 词、最终回复 ≤100 词」的硬性限制。
ablation 测试显示该约束使整体智能下降 3%。

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix: 修复测试中 reasoning_content 类型断言

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

src/bootstrap/state.ts

@@ -235,11 +235,6 @@ type State = {
   // microcompact is first enabled, keep sending the header so mid-session
   // GrowthBook/settings toggles don't bust the prompt cache.
   cacheEditingHeaderLatched: boolean | null
-  // Sticky-on latch for clearing thinking from prior tool loops. Triggered
-  // when >1h since last API call (confirmed cache miss — no cache-hit
-  // benefit to keeping thinking). Once latched, stays on so the newly-warmed
-  // thinking-cleared cache isn't busted by flipping back to keep:'all'.
-  thinkingClearLatched: boolean | null
   // Current prompt ID (UUID) correlating a user prompt with subsequent OTel events
   promptId: string | null
   // Last API requestId for the main conversation chain (not subagents).
@@ -414,7 +409,6 @@ function getInitialState(): State {
     afkModeHeaderLatched: null,
     fastModeHeaderLatched: null,
     cacheEditingHeaderLatched: null,
-    thinkingClearLatched: null,
     // Current prompt ID
     promptId: null,
     lastMainRequestId: undefined,
@@ -1729,14 +1723,6 @@ export function setCacheEditingHeaderLatched(v: boolean): void {
   STATE.cacheEditingHeaderLatched = v
 }
 
-export function getThinkingClearLatched(): boolean | null {
-  return STATE.thinkingClearLatched
-}
-
-export function setThinkingClearLatched(v: boolean): void {
-  STATE.thinkingClearLatched = v
-}
-
 /**
  * Reset beta header latches to null. Called on /clear and /compact so a
  * fresh conversation gets fresh header evaluation.
@@ -1745,7 +1731,6 @@ export function clearBetaHeaderLatches(): void {
   STATE.afkModeHeaderLatched = null
   STATE.fastModeHeaderLatched = null
   STATE.cacheEditingHeaderLatched = null
-  STATE.thinkingClearLatched = null
 }
 
 export function getPromptId(): string | null {

src/constants/prompts.ts

@@ -614,17 +614,6 @@ ${CYBER_RISK_INSTRUCTION}`,
       'summarize_tool_results',
       () => SUMMARIZE_TOOL_RESULTS_SECTION,
     ),
-    // Numeric length anchors — research shows ~1.2% output token reduction vs
-    // qualitative "be concise". Ant-only to measure quality impact first.
-    ...(process.env.USER_TYPE === 'ant'
-      ? [
-          systemPromptSection(
-            'numeric_length_anchors',
-            () =>
-              'Length limits: keep text between tool calls to \u226425 words. Keep final responses to \u2264100 words unless the task requires more detail.',
-          ),
-        ]
-      : []),
     ...(feature('TOKEN_BUDGET')
       ? [
           // Cached unconditionally — the "When the user specifies..." phrasing

src/services/api/claude.ts

@@ -124,14 +124,12 @@ import {
   getPromptCache1hAllowlist,
   getPromptCache1hEligible,
   getSessionId,
-  getThinkingClearLatched,
   setAfkModeHeaderLatched,
   setCacheEditingHeaderLatched,
   setFastModeHeaderLatched,
   setLastMainRequestId,
   setPromptCache1hAllowlist,
   setPromptCache1hEligible,
-  setThinkingClearLatched,
 } from 'src/bootstrap/state.js'
 import {
   AFK_MODE_BETA_HEADER,
@@ -1492,20 +1490,6 @@ async function* queryModel(
     }
   }
 
-  // Only latch from agentic queries so a classifier call doesn't flip the
-  // main thread's context_management mid-turn.
-  let thinkingClearLatched = getThinkingClearLatched() === true
-  if (!thinkingClearLatched && isAgenticQuery) {
-    const lastCompletion = getLastApiCompletionTimestamp()
-    if (
-      lastCompletion !== null &&
-      Date.now() - lastCompletion > CACHE_TTL_1HOUR_MS
-    ) {
-      thinkingClearLatched = true
-      setThinkingClearLatched(true)
-    }
-  }
-
   const effort = resolveAppliedEffort(options.model, options.effortValue)
 
   if (feature('PROMPT_CACHE_BREAK_DETECTION')) {
@@ -1684,7 +1668,7 @@ async function* queryModel(
     const contextManagement = getAPIContextManagement({
       hasThinking,
       isRedactThinkingActive: betasParams.includes(REDACT_THINKING_BETA_HEADER),
-      clearAllThinking: thinkingClearLatched,
+      clearAllThinking: false,
     })
 
     const enablePromptCaching =

src/services/api/openai/__tests__/thinking.test.ts

@@ -100,16 +100,28 @@ describe('isOpenAIThinkingEnabled', () => {
       expect(isOpenAIThinkingEnabled('TokenService/deepseek-v3.2')).toBe(true)
     })
 
-    test('returns false when model name is "deepseek-chat"', () => {
-      expect(isOpenAIThinkingEnabled('deepseek-chat')).toBe(false)
+    test('returns true when model name is "deepseek-chat"', () => {
+      expect(isOpenAIThinkingEnabled('deepseek-chat')).toBe(true)
     })
 
-    test('returns false when model name is "deepseek-v3"', () => {
-      expect(isOpenAIThinkingEnabled('deepseek-v3')).toBe(false)
+    test('returns true when model name is "deepseek-v3"', () => {
+      expect(isOpenAIThinkingEnabled('deepseek-v3')).toBe(true)
     })
 
-    test('returns false when model name contains "deepseek" but not "reasoner" or "v3.2"', () => {
-      expect(isOpenAIThinkingEnabled('deepseek-coder')).toBe(false)
+    test('returns true when model name is "deepseek-v4"', () => {
+      expect(isOpenAIThinkingEnabled('deepseek-v4')).toBe(true)
+    })
+
+    test('returns true when model name is "deepseek-v4-pro"', () => {
+      expect(isOpenAIThinkingEnabled('deepseek-v4-pro')).toBe(true)
+    })
+
+    test('returns true when model name is "deepseek-r1"', () => {
+      expect(isOpenAIThinkingEnabled('deepseek-r1')).toBe(true)
+    })
+
+    test('returns true when model name contains "deepseek"', () => {
+      expect(isOpenAIThinkingEnabled('deepseek-coder')).toBe(true)
     })
 
     test('returns false when model name is "gpt-4o"', () => {
@@ -126,6 +138,7 @@ describe('isOpenAIThinkingEnabled', () => {
       process.env.OPENAI_ENABLE_THINKING = '1'
       expect(isOpenAIThinkingEnabled('gpt-4o')).toBe(true)
       expect(isOpenAIThinkingEnabled('deepseek-v3')).toBe(true)
+      expect(isOpenAIThinkingEnabled('qwen-3')).toBe(true)
     })
 
     test('OPENAI_ENABLE_THINKING=false disables thinking even for deepseek-reasoner', () => {

src/services/api/openai/requestBody.ts

@@ -25,9 +25,9 @@ export function isOpenAIThinkingEnabled(model: string): boolean {
   if (isEnvDefinedFalsy(process.env.OPENAI_ENABLE_THINKING)) return false
   // Explicit enable
   if (isEnvTruthy(process.env.OPENAI_ENABLE_THINKING)) return true
-  // Auto-detect from model name (deepseek-reasoner and DeepSeek-V3.2 support thinking mode)
+  // Auto-detect from model name (all DeepSeek models support thinking mode)
   const modelLower = model.toLowerCase()
-  return modelLower.includes('deepseek-reasoner') || modelLower.includes('deepseek-v3.2')
+  return modelLower.includes('deepseek')
 }
 
 /**

src/services/api/src/bootstrap/state.ts

@@ -6,14 +6,12 @@ export type getFastModeHeaderLatched = any;
 export type getLastApiCompletionTimestamp = any;
 export type getPromptCache1hAllowlist = any;
 export type getPromptCache1hEligible = any;
-export type getThinkingClearLatched = any;
 export type setAfkModeHeaderLatched = any;
 export type setCacheEditingHeaderLatched = any;
 export type setFastModeHeaderLatched = any;
 export type setLastMainRequestId = any;
 export type setPromptCache1hAllowlist = any;
 export type setPromptCache1hEligible = any;
-export type setThinkingClearLatched = any;
 export type addToTotalDurationState = any;
 export type consumePostCompaction = any;
 export type getIsNonInteractiveSession = any;

src/utils/effort.ts

@@ -348,13 +348,13 @@ export function getDefaultEffortForModel(
     model.toLowerCase().includes('opus-4-6')
   ) {
     if (isProSubscriber()) {
-      return 'medium'
+      return 'high'
     }
     if (
       getOpusDefaultEffortConfig().enabled &&
       (isMaxSubscriber() || isTeamSubscriber())
     ) {
-      return 'medium'
+      return 'high'
     }
   }