fix: walk back round-1/2 conclusions — the cause was think=false all along

Seth asked "was this with think=false?" Yes — and that was the only
question that mattered. Everything I concluded in round 1 and round 2
was wrong.

Actual cause, isolated in round 3:
- At identical message state, gemma4:26b with think=false returns
  eval=4 (silent stop); with think unset or think=true, returns
  eval=165 and emits the correct tool call.
- Original round-1 write_file harness + think unset: 26B passes in
  8 iters, 20s. No mitigations needed.
- 31B dense and qwen3-coder:30b tolerate think=false; 26B MoE does not.

Red herrings (kept on-record in the bakeoff doc, not silently erased):
- Round 1: "write_file tool-call argument size" — wrong
- Round 2a: refuted the arg-size theory but for the wrong reason
  (still failed because think=false was still set)
- Round 2b: "cumulative tool-response context size" — truncating
  did make 26B pass, but by coincidence. Shorter context at the
  decision turn dodged the think=false side effect.

Why the existing "always think:false" guidance was misleading:
it was derived from AI_Visualizer (single-turn JSON pipelines) where
thinking tokens do eat num_predict invisibly. In multi-turn
tool-calling agents the channels are separate and the flag has a
different effect — catastrophic on 26B specifically.

Doc updates:
- GOTCHAS: replaced the 26B entry with the actual cause; scoped the
  original "Thinking Mode Eats Context" entry to single-turn pipelines
- SYNTHESIS: split the "Mandatory Ollama Settings" block into
  single-turn vs multi-turn variants; updated anti-patterns and
  quick-start checklist
- CORPUS_cli_coding_agent.md: revised pointer and config template
- docs/reference/bakeoff-2026-04-18.md: added Round 3 section with
  the correction notice at the top of the file and full diagnostic
  methodology

New artifacts: harness_no_think_flag.py, harness_write_no_think.py,
and 4 new log files demonstrating all three models pass when think
is left at default.

scripts/bakeoff/runs_patch/gemma4-26b-writefile-no-think/log.json

@@ -0,0 +1,83 @@
+{
+  "model": "gemma4:26b",
+  "edit_tool": "write_file",
+  "think_setting": "unset (default)",
+  "turns": [
+    {
+      "iteration": 1,
+      "elapsed_s": 8.52,
+      "content": "",
+      "prompt_eval_count": 369,
+      "eval_count": 55,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 2,
+      "elapsed_s": 1.07,
+      "content": "",
+      "prompt_eval_count": 510,
+      "eval_count": 46,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 3,
+      "elapsed_s": 0.75,
+      "content": "",
+      "prompt_eval_count": 691,
+      "eval_count": 19,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 4,
+      "elapsed_s": 1.05,
+      "content": "",
+      "prompt_eval_count": 1140,
+      "eval_count": 27,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 5,
+      "elapsed_s": 1.19,
+      "content": "",
+      "prompt_eval_count": 1786,
+      "eval_count": 23,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 6,
+      "elapsed_s": 4.87,
+      "content": "",
+      "prompt_eval_count": 2075,
+      "eval_count": 330,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 7,
+      "elapsed_s": 0.99,
+      "content": "",
+      "prompt_eval_count": 2422,
+      "eval_count": 26,
+      "tool_calls_count": 1
+    },
+    {
+      "iteration": 8,
+      "elapsed_s": 0.96,
+      "content": "The `median` function was updated to correctly average the two middle elements when the input list has an even number of elements.",
+      "prompt_eval_count": 2641,
+      "eval_count": 30,
+      "tool_calls_count": 0
+    }
+  ],
+  "started_at": 1776550159.4912357,
+  "final": {
+    "halt_reason": "no_tool_calls",
+    "tests_pass": true,
+    "iterations_used": 8,
+    "tool_call_counts": {
+      "read_file": 2,
+      "write_file": 1,
+      "run_bash": 4
+    },
+    "wall_clock_s": 20.58
+  }
+}