From 98ea4c4f545b3e8a81cc412ece0191c154012dba Mon Sep 17 00:00:00 2001
From: Devika Verma <devikaverma11@gmail.com>
Date: Mon, 17 Aug 2026 18:01:38 +0530
Subject: [PATCH] render_history: label ask_user replies 'reply', not 'turn N'

A turn is a message the user sent on their own; labelling an answer as
one reads as a spontaneous statement - stronger evidence than it is.
Turn numbering now counts real messages only.
---
 coworker/reviewer.py       | 15 +++++++++++----
 tests/test_auto_approve.py | 15 +++++++++++++++
 2 files changed, 26 insertions(+), 4 deletions(-)

diff --git a/coworker/reviewer.py b/coworker/reviewer.py
index 47b6d27e..0767e200 100644
--- a/coworker/reviewer.py
+++ b/coworker/reviewer.py
@@ -202,16 +202,23 @@ def render_history(user_messages: list[dict[str, Any]]) -> str:
     """The EARLIER-IN-THIS-SESSION block: the user's own words, mechanically extracted,
     clipped hard, with `ask_user` replies tagged as replies (§8.2). `user_messages` is a
     list of {"text": str, "is_reply": bool} in chronological order, current turn excluded.
-    """
+
+    Replies are labelled `reply`, never `turn N`: a "turn" is a message the user sent on
+    their own, and labelling an answer as one would read as a spontaneous statement —
+    stronger evidence than it is. Turn numbering counts real messages only."""
     if not user_messages:
         return ""
     lines = ["EARLIER IN THIS SESSION (the user's own words, verbatim)"]
-    for i, msg in enumerate(user_messages, start=1):
+    turn = 0
+    for msg in user_messages:
         text = clip_message(str(msg.get("text", "")))
         if not text:
             continue
-        tag = "  [reply to a question the agent asked]" if msg.get("is_reply") else ""
-        lines.append(f"  turn {i}  {text}{tag}")
+        if msg.get("is_reply"):
+            lines.append(f"  reply   {text}  [reply to a question the agent asked]")
+        else:
+            turn += 1
+            lines.append(f"  turn {turn}  {text}")
     return "\n".join(lines) if len(lines) > 1 else ""
 
 
diff --git a/tests/test_auto_approve.py b/tests/test_auto_approve.py
index 2c007540..f836646b 100644
--- a/tests/test_auto_approve.py
+++ b/tests/test_auto_approve.py
@@ -158,6 +158,21 @@ def test_reply_tag_is_rendered():
     assert "[reply to a question the agent asked]" in rendered
 
 
+def test_replies_are_labelled_reply_and_turn_numbering_skips_them():
+    # A reply is not a "turn": labelling it as one would read as a spontaneous statement.
+    rendered = reviewer_mod.render_history(
+        [
+            {"text": "fix the tests"},
+            {"text": "yes", "is_reply": True},
+            {"text": "update the changelog"},
+        ]
+    )
+    lines = rendered.splitlines()[1:]
+    assert lines[0].startswith("  turn 1  fix the tests")
+    assert lines[1].startswith("  reply   yes") and "turn" not in lines[1]
+    assert lines[2].startswith("  turn 2  update the changelog")  # numbering skipped the reply
+
+
 # -- ask_user answers reach the reviewer history (§8.2 reply capture) --------------