Latch cache beta headers from requests

Author: SqlRush

docs/claude-code-go-rewrite-plan.md

@@ -132,6 +132,8 @@ test/parity/                 # golden tests against TS/official behavior
 - 不启用工具时的 prompt -> streaming -> assistant 输出可跑通。
 - API request golden 能比对 headers/body 中关键字段。
 
+- 本轮补充：Anthropic client 构造请求时会把显式 beta headers 与请求内容推导出的 beta headers 合并去重；当 messages/system content 或 tool definitions 中出现 `cache_control`/`cache_reference`，以及 messages 中出现 `cache_edits` payload 时，会自动带上 prompt-caching scope 与 cache-editing beta，避免 prompt-cache/cache-editing 请求依赖外部环境变量手动配置。
+
 ### M4: Tool framework、permissions、sandbox
 
 产出：

docs/first-second-parity-audit.md

@@ -93,7 +93,7 @@ Anthropic API 和 conversation：
 - Model registry now uses current Claude model IDs and aliases from the source snapshot, including Sonnet 4.6, Opus 4.6, Haiku 4.5, canonical-name rendering, and `[1m]` context variants.
 - OAuth support now includes production OAuth config, scope parsing, Claude.ai scope detection, auth URL construction, PKCE verifier/challenge, state generation, and expiry checks.
 - Session/history support now includes CC-compatible prompt history references, pasted text/image placeholder parsing, paste-cache hashing and retrieval, `history.jsonl` append/load, current-session-first up-arrow ordering, ctrl+r-style deduped timestamped history, `CLAUDE_CODE_SKIP_PROMPT_HISTORY`, remote session event pagination helpers, lenient transcript loading, legacy progress parent-bridge recovery, compact-boundary pruning, snip removal/relink replay, metadata entry collection, leaf UUID calculation, conversation-chain reconstruction, orphaned parallel tool-result recovery, content-replacement record loading/reconstruction, tombstone metadata delete/relink replay, and tombstone-style transcript message removal with a size guard.
-- Anthropic API layer now covers streaming accumulation, usage update/accumulation semantics, non-streaming max token cap, thinking-budget adjustment, retry/backoff with `Retry-After` and `x-should-retry`, context-overflow `max_tokens` retry adjustment, beta-header dedupe, custom request headers, basic prompt cache breakpoint/cache-reference/cache-edits placement, prompt dump JSONL capture for init/new user messages/non-streaming responses/stream chunks, and CC-compatible USD cost calculation for known Claude models including cache read/write, web search requests, and Opus 4.6 fast-tier pricing.
+- Anthropic API layer now covers streaming accumulation, usage update/accumulation semantics, non-streaming max token cap, thinking-budget adjustment, retry/backoff with `Retry-After` and `x-should-retry`, context-overflow `max_tokens` retry adjustment, beta-header dedupe, request-content beta latching for prompt-cache/cache-editing payloads, custom request headers, basic prompt cache breakpoint/cache-reference/cache-edits placement, prompt dump JSONL capture for init/new user messages/non-streaming responses/stream chunks, and CC-compatible USD cost calculation for known Claude models including cache read/write, web search requests, and Opus 4.6 fast-tier pricing.
 - Tool runtime now includes concurrency partitioning, ordered concurrent execution, interrupt behavior/defaults, max result size metadata, oversized result persistence, pre/post/permission-denied/permission-request hook dispatch, settings and local-plugin command-backed/HTTP hook execution for synchronous tool and conversation lifecycle hooks, hook-driven input updates/blocking/permission-request allow-deny, prompt/compact context injection, executor hook phase progress events, lifecycle progress events, retry/fallback model breadcrumbs, final-result/structured-error model attempt lists, structured API error status/type/request-id fields, headless stream-json progress events, snake_case token warning NDJSON payloads, lightweight compact NDJSON metadata, and pre-call cancellation checks.
 - Conversation runner can now use streaming clients, aggregate stream events into assistant messages, run tool calls through the orchestrator, preserve transcript append behavior, and apply CC-style per-message aggregate tool-result budget replacement before API requests with persisted replacement records for resume.
 
@@ -106,7 +106,7 @@ The following items remain incomplete and must not be treated as done:
 - Interactive permission prompt flow: REPL dialogs, bridge/channel/swarm permission relays, user feedback images, prompt race handling, cancellation.
 - Full filesystem permission parity gaps that remain: skill-scope allow suggestions, complete local/local-jsx slash-command execution wiring, command permission UI/SDK surfacing, forked/remote/MCP/plugin/bundled SkillTool behavior, skill prompt shell injection, complete auto-memory override policy, and deeper platform-specific Windows/WSL bypass handling.
 - Full tool execution parity gaps that remain: complete hook runtime policy beyond the current settings command hook path, MCP elicitation, complete SDK control event surface beyond current progress NDJSON, mid-call cancellation for concrete tools, background task behavior, telemetry, schema-not-sent hints, and concrete tool-specific semantics.
-- Complete Anthropic API parity gaps that remain: full dynamic beta-header latching, ant-only dump gating and `/issue` integration, cost tracker/session restore integration, streaming-to-non-streaming fallback, gateway/proxy-specific headers, first-party/Bedrock/Vertex/Foundry client setup, OAuth refresh in retry loop, fast-mode retry/cooldown semantics, persistent unattended retry heartbeats, full prompt-cache editing lifecycle, and provider-specific cache behavior.
+- Complete Anthropic API parity gaps that remain: dynamic beta-header latching beyond prompt-cache/cache-editing payloads, ant-only dump gating and `/issue` integration, cost tracker/session restore integration, streaming-to-non-streaming fallback, gateway/proxy-specific headers, first-party/Bedrock/Vertex/Foundry client setup, OAuth refresh in retry loop, fast-mode retry/cooldown semantics, persistent unattended retry heartbeats, full prompt-cache editing lifecycle, and provider-specific cache behavior.
 - Full conversation/query loop: stop hooks, compact/auto-compact, token budget escalation, resume, SDK JSON/NDJSON control events, status updates, rate-limit handling, model switch breadcrumbs, side questions.
 - Full settings parity gaps that remain: complete Zod-equivalent validation messages, MDM/HKCU settings, managed-settings drop-in loading, settings cache/change detector, schema generation, plugin-only customization enforcement, all strict marketplace validations, and live reload/app-state sync.
 - Full session/history parity gaps that remain: large-file optimized transcript loading, preserved-segment edge cases beyond current relink/prune support, content-replacement feature-flag/runtime override and inherited subagent gap-fill details, complete async prompt-history lifecycle parity beyond current lock/buffer/undo paths, full pasted-image processing/runtime integration beyond current prompt image-cache/image-block/metadata path, remaining remote-history edge cases, sidechain/subagent transcript layout, and all session metadata entry types.

internal/api/anthropic/betas.go

@@ -1,6 +1,10 @@
 package anthropic
 
-import "strings"
+import (
+	"strings"
+
+	"ccgo/internal/contracts"
+)
 
 const (
 	PromptCachingScopeBetaHeader = "prompt-caching-scope-2024-07-31"
@@ -32,3 +36,58 @@ func MergeBetaHeaders(groups ...[]string) []string {
 func BetaHeaderValue(headers []string) string {
 	return strings.Join(MergeBetaHeaders(headers), ",")
 }
+
+func DynamicBetaHeaders(payload any) []string {
+	request, ok := payload.(Request)
+	if !ok {
+		return nil
+	}
+	var betas []string
+	if requestUsesPromptCaching(request) {
+		betas = append(betas, PromptCachingScopeBetaHeader)
+	}
+	if requestUsesCacheEditing(request) {
+		betas = append(betas, CacheEditingBetaHeader)
+	}
+	return betas
+}
+
+func requestUsesPromptCaching(request Request) bool {
+	for _, msg := range request.Messages {
+		if contentBlocksUsePromptCaching(msg.Content) {
+			return true
+		}
+	}
+	if systemBlocks, ok := request.System.([]contracts.ContentBlock); ok && contentBlocksUsePromptCaching(systemBlocks) {
+		return true
+	}
+	for _, tool := range request.Tools {
+		if tool.CacheControl != nil {
+			return true
+		}
+	}
+	return false
+}
+
+func requestUsesCacheEditing(request Request) bool {
+	for _, msg := range request.Messages {
+		for _, block := range msg.Content {
+			if block.Type == contracts.ContentCacheEdits || len(block.Edits) > 0 {
+				return true
+			}
+		}
+	}
+	return false
+}
+
+func contentBlocksUsePromptCaching(blocks []contracts.ContentBlock) bool {
+	for _, block := range blocks {
+		if block.CacheControl != nil || strings.TrimSpace(block.CacheReference) != "" {
+			return true
+		}
+		if block.Type == contracts.ContentCacheEdits || len(block.Edits) > 0 {
+			return true
+		}
+	}
+	return false
+}

internal/api/anthropic/client.go

@@ -289,8 +289,9 @@ func (c *Client) newJSONRequestWithDump(ctx context.Context, path string, payloa
 	req.Header.Set("content-type", "application/json")
 	req.Header.Set("accept", "application/json")
 	req.Header.Set("anthropic-version", c.version())
-	if len(c.Beta) > 0 {
-		req.Header.Set("anthropic-beta", BetaHeaderValue(c.Beta))
+	betas := MergeBetaHeaders(c.Beta, DynamicBetaHeaders(payload))
+	if len(betas) > 0 {
+		req.Header.Set("anthropic-beta", BetaHeaderValue(betas))
 	}
 	for key, values := range c.Headers {
 		for _, value := range values {

internal/api/anthropic/client_test.go

@@ -332,6 +332,43 @@ func TestBetaHeadersDedupedAndCustomHeadersApplied(t *testing.T) {
 	}
 }
 
+func TestCreateMessageAddsDynamicCacheBetaHeaders(t *testing.T) {
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if got := r.Header.Get("anthropic-beta"); got != "one,prompt-caching-scope-2024-07-31,cache-editing-2025-01-24" {
+			t.Fatalf("anthropic-beta = %q", got)
+		}
+		w.Header().Set("content-type", "application/json")
+		_, _ = w.Write([]byte(`{"id":"msg_1","type":"message","role":"assistant","model":"sonnet","content":[]}`))
+	}))
+	defer server.Close()
+
+	client := NewClient(WithBaseURL(server.URL), WithBeta("one", PromptCachingScopeBetaHeader))
+	_, err := client.CreateMessage(context.Background(), Request{
+		Model:     "sonnet",
+		MaxTokens: 32,
+		Messages: []contracts.APIMessage{{
+			Role: "user",
+			Content: []contracts.ContentBlock{
+				{
+					Type:         contracts.ContentText,
+					Text:         "hello",
+					CacheControl: &contracts.CacheControl{Type: "ephemeral", Scope: "global"},
+				},
+				{
+					Type: contracts.ContentCacheEdits,
+					Edits: []contracts.CacheEdit{{
+						Type:           "delete",
+						CacheReference: "toolu_old",
+					}},
+				},
+			},
+		}},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+}
+
 func TestAddCacheBreakpoints(t *testing.T) {
 	messages := []contracts.APIMessage{
 		{Role: "user", Content: []contracts.ContentBlock{{Type: contracts.ContentToolResult, ToolUseID: "toolu_1"}}},