From f14ad0c69394f4187c8eff713b836545fcf9f4d6 Mon Sep 17 00:00:00 2001
From: song <130889452+mrsxs@users.noreply.github.com>
Date: Tue, 21 Apr 2026 12:07:14 +0800
Subject: [PATCH] fix(llmcore): preserve thinking block signature in streaming
 SSE parser (#123)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Anthropic's extended thinking streaming protocol emits two delta types
for a single thinking block: `thinking_delta` (the textual reasoning)
and `signature_delta` (a base64 HMAC tag appended at the end of the
block). Both must be accumulated into the same `content_block`.

Current code only handles `thinking_delta`, so `signature_delta` events
are silently dropped. When the assistant's reply (with thinking) is
echoed back on the next turn, Anthropic's server validates the
signature and rejects the request with 400:

    "Invalid `signature` in `thinking` block"

Downstream effects observed in production (via sub2api relay logs):
- Every request with history triggers a 400 signature error
- The relay strips thinking blocks and retries, which changes the
  cache prefix and invalidates prompt caching, forcing a full rebuild
  of cache_creation_tokens (~20k-30k per affected request)
- Measured in a 5h window: 5/25 requests suffered cache invalidation,
  accounting for 53.5% of total spend that was otherwise avoidable

Fix:
1. Initialize `current_block` with an empty `signature` field when a
   thinking block starts, so the dict shape matches Anthropic's spec
   (`{type, thinking, signature}`).
2. Handle `signature_delta` events by appending `delta.signature` to
   `current_block["signature"]`. Using `+=` (rather than assignment)
   mirrors how `thinking_delta` is accumulated and is robust against
   future chunked signatures.

No behavior change for clients that disable extended thinking, or for
upstreams that don't emit `signature_delta`. For `tool_use` threads
that require valid thinking signatures to preserve reasoning context,
this fix is required — the previous behavior silently corrupted them.

Verification:
- Before fix: upstream returns 400 + retry; cache_creation_tokens
  spike to ~25k on every 4th-5th request in a conversation
- After fix: upstream accepts the first attempt; cache_read_tokens
  dominate, cache_creation_tokens only appear on the first request
  of a fresh 5m prompt-cache window
---
 llmcore.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/llmcore.py b/llmcore.py
index c8d9970..604f97c 100644
--- a/llmcore.py
+++ b/llmcore.py
@@ -115,7 +115,7 @@ def _parse_claude_sse(resp_lines):
         elif evt_type == "content_block_start":
             block = evt.get("content_block", {})
             if block.get("type") == "text": current_block = {"type": "text", "text": ""}
-            elif block.get("type") == "thinking": current_block = {"type": "thinking", "thinking": ""}
+            elif block.get("type") == "thinking": current_block = {"type": "thinking", "thinking": "", "signature": ""}
             elif block.get("type") == "tool_use":
                 current_block = {"type": "tool_use", "id": block.get("id", ""), "name": block.get("name", ""), "input": {}}
                 tool_json_buf = ""
@@ -127,6 +127,9 @@ def _parse_claude_sse(resp_lines):
                 if text: yield text
             elif delta.get("type") == "thinking_delta":
                 if current_block and current_block.get("type") == "thinking": current_block["thinking"] += delta.get("thinking", "")
+            elif delta.get("type") == "signature_delta":
+                if current_block and current_block.get("type") == "thinking":
+                    current_block["signature"] = current_block.get("signature", "") + delta.get("signature", "")
             elif delta.get("type") == "input_json_delta": tool_json_buf += delta.get("partial_json", "")
         elif evt_type == "content_block_stop":
             if current_block: