From 63a7810b52341fcf676bec91d2153f604c80d168 Mon Sep 17 00:00:00 2001
From: Steven Chong <25894545+teamchong@users.noreply.github.com>
Date: Tue, 26 May 2026 11:05:59 -0400
Subject: [PATCH] fix(history): preserve git status output as text

---
 src/core/history.ts   | 65 +++++++++++++++++++++++++++++++++++++++----
 tests/history.test.ts | 35 +++++++++++++++++++++++
 2 files changed, 95 insertions(+), 5 deletions(-)

diff --git a/src/core/history.ts b/src/core/history.ts
index 4aa8d9c..d3fdd17 100644
--- a/src/core/history.ts
+++ b/src/core/history.ts
@@ -113,6 +113,48 @@ export interface HistoryCollapseInfo {
   droppedCodepoints: Map<number, number>;
 }
 
+const MAX_VERBATIM_GIT_STATUS_CHARS = 8000;
+
+function toolResultTextContent(block: ToolResultBlock): string {
+  const inner = block.content;
+  if (typeof inner === 'string') return inner;
+  if (!Array.isArray(inner)) return '';
+  const parts: string[] = [];
+  for (const sub of inner) {
+    if (sub && (sub as TextBlock).type === 'text') parts.push((sub as TextBlock).text);
+  }
+  return parts.join('\n');
+}
+
+function isGitStatusShortOutput(text: string): boolean {
+  if (!text || text.length > MAX_VERBATIM_GIT_STATUS_CHARS) return false;
+  const lines = text.split(/\r?\n/).filter((line) => line.length > 0);
+  if (lines.length === 0) return false;
+  return lines.every((line) => /^(?:[ MADRCUT?!]{2}|[?!]{2}) .+/.test(line));
+}
+
+function collectVerbatimGitStatusOutputs(
+  messages: Message[],
+  upToExclusive: number,
+): string[] {
+  const out: string[] = [];
+  let used = 0;
+  const limit = Math.min(upToExclusive, messages.length);
+  for (let i = 0; i < limit; i++) {
+    const msg = messages[i]!;
+    if (!Array.isArray(msg.content)) continue;
+    for (const blk of msg.content) {
+      if (!blk || (blk as { type?: string }).type !== 'tool_result') continue;
+      const text = toolResultTextContent(blk as ToolResultBlock);
+      if (!isGitStatusShortOutput(text)) continue;
+      if (used + text.length > MAX_VERBATIM_GIT_STATUS_CHARS) return out;
+      out.push(text);
+      used += text.length;
+    }
+  }
+  return out;
+}
+
 /**
  * Compute the largest index `i*` in `[0..messages.length - keepTail - 1]`
  * such that all `tool_use_id`s issued by assistant turns in `[0..i*]` are
@@ -367,13 +409,26 @@ export async function collapseHistory(
     }
   }
   // Build the synthetic user message.
+  const verbatimGitStatusOutputs = collectVerbatimGitStatusOutputs(messages, collapseLen);
+  const verbatimBlocks: TextBlock[] = verbatimGitStatusOutputs.length > 0
+    ? [{
+        type: 'text',
+        text: [
+          '[Exact git status --short output preserved as text. Use this block as the source of truth over the rendered history image.]',
+          ...verbatimGitStatusOutputs.map((out, i) => `--- git status --short #${i + 1} ---\n${out}`),
+          '[End exact git status --short output.]',
+        ].join('\n\n'),
+      }]
+    : [];
+  const syntheticContent: ContentBlock[] = [
+    { type: 'text', text: '[Earlier in this conversation:]' },
+    ...imageBlocks,
+    ...verbatimBlocks,
+    { type: 'text', text: '[End of earlier context.]' },
+  ];
   const syntheticUser: Message = {
     role: 'user',
-    content: [
-      { type: 'text', text: '[Earlier in this conversation:]' },
-      ...imageBlocks,
-      { type: 'text', text: '[End of earlier context.]' },
-    ],
+    content: syntheticContent,
   };
   const tail = messages.slice(collapseLen);
   info.collapsedTurns = collapseLen;
diff --git a/tests/history.test.ts b/tests/history.test.ts
index 0df75d7..4ee1be6 100644
--- a/tests/history.test.ts
+++ b/tests/history.test.ts
@@ -312,6 +312,41 @@ describe('collapseHistory', () => {
     expect(out[2]).toBe(msgs[11]);
   });
 
+  it('preserves git status --short tool output as exact text alongside history images', async () => {
+    const status = [
+      'M  package-lock.json',
+      'M  package.json',
+      'M  packages/capnweb-typecheck/DESIGN.md',
+      'M  packages/capnweb-typecheck/src/index.ts',
+      ' M packages/capnweb-typecheck/src/plugin.ts',
+      'M  tsdown.config.ts',
+      '?? handoff.md',
+      '?? packages/capnweb-typecheck/src/transform/',
+    ].join('\n');
+    const msgs: Message[] = [
+      usr('run git status'),
+      asst([{ type: 'tool_use', id: 'bash-1', name: 'bash', input: { command: 'git status --short' } }]),
+      usr([{ type: 'tool_result', tool_use_id: 'bash-1', content: status }]),
+      asst('noted'),
+    ];
+
+    const { messages: out, info } = await collapseHistory(msgs, () => true, {
+      keepTail: 0,
+      minCollapsePrefix: 1,
+      collapseChunk: 0,
+    });
+
+    expect(info.reason).toBe(undefined);
+    const content = out[0]!.content as Array<Record<string, unknown>>;
+    expect(content.some((c) => c.type === 'image')).toBe(true);
+    const exact = content.find(
+      (c) => c.type === 'text' && typeof c.text === 'string' && c.text.includes('source of truth'),
+    );
+    expect(exact).toBeTruthy();
+    expect((exact!.text as string)).toContain(status);
+    for (const line of status.split('\n')) expect((exact!.text as string)).toContain(line);
+  });
+
   it('preserves a tool_use sequence that straddles the live-tail boundary', async () => {
     // 14 turns: 10 closed turns, then an open tool_use at index 10 that closes at index 12.
     // Per-turn body bumped to 4200 chars so the row-aware gate (numCols=1) clears