From cf3443796ce23fba3f7495b64a2cde2a8cd5afd7 Mon Sep 17 00:00:00 2001
From: Vincent Koc <25068+vincentkoc@users.noreply.github.com>
Date: Fri, 10 Jul 2026 20:48:28 -0700
Subject: [PATCH] fix(testbox): sync source on every lease reuse

---
 docs/reference/test.md                       |  7 +++---
 scripts/crabbox-wrapper.mjs                  | 16 --------------
 scripts/testbox-lease-freshness.mjs          | 16 +++-----------
 test/scripts/testbox-lease-freshness.test.ts | 23 +-------------------
 4 files changed, 7 insertions(+), 55 deletions(-)

diff --git a/docs/reference/test.md b/docs/reference/test.md
index 46ed426089d6..b44abcda821f 100644
--- a/docs/reference/test.md
+++ b/docs/reference/test.md
@@ -28,10 +28,9 @@ After the first successful reuse, the wrapper records the lease's base,
 dependency, and Testbox workflow fingerprint under `.crabbox/testbox-leases/`.
 Source-only edits keep reusing the warmed box. A changed merge base, lockfile,
 package-manager input, wrapper, or Testbox workflow fails closed and requires a
-fresh lease. `OPENCLAW_TESTBOX_ALLOW_STALE=1` is only for intentional
-diagnostics, not release proof. Repeating a targeted run from the exact clean
-HEAD that already succeeded skips source sync; changed gates, a new commit, or
-a dirty edit sync normally.
+fresh lease. Every run still syncs the current checkout.
+`OPENCLAW_TESTBOX_ALLOW_STALE=1` is only for intentional diagnostics, not
+release proof.
 
 Local test commands below are for human workflows or an explicit agent fallback
 requested by the user. Remote-provider unavailability must be reported; it is
diff --git a/scripts/crabbox-wrapper.mjs b/scripts/crabbox-wrapper.mjs
index 1ec4cda43309..b73310ed0edf 100755
--- a/scripts/crabbox-wrapper.mjs
+++ b/scripts/crabbox-wrapper.mjs
@@ -3184,16 +3184,6 @@ function injectFullCheckoutLeaseReclaim(commandArgs) {
   return normalizedArgs;
 }
 
-function injectReusableTestboxNoSync(commandArgs) {
-  if (hasOption(commandArgs, "--no-sync")) {
-    return commandArgs;
-  }
-  const normalizedArgs = [...commandArgs];
-  const { optionEnd } = runCommandBounds(normalizedArgs);
-  normalizedArgs.splice(optionEnd, 0, "--no-sync");
-  return normalizedArgs;
-}
-
 function injectRemoteTestboxCi(commandArgs, providerName) {
   if (commandArgs[0] !== "run" || canonicalProviderName(providerName) !== "blacksmith-testbox") {
     return commandArgs;
@@ -3299,12 +3289,6 @@ try {
     provider: canonicalProvider,
     repoRoot,
   });
-  if (testboxLeaseFreshness?.skipSync && !isChangedGateCommand(runCommandArgs(normalizedArgs))) {
-    normalizedArgs = injectReusableTestboxNoSync(normalizedArgs);
-    console.error(
-      `[crabbox] Testbox ${optionValue(normalizedArgs, "--id")} already has clean HEAD ${testboxLeaseFreshness.current.headSha}; skipping source sync`,
-    );
-  }
 } catch (error) {
   console.error(`[crabbox] ${error instanceof Error ? error.message : String(error)}`);
   process.exit(2);
diff --git a/scripts/testbox-lease-freshness.mjs b/scripts/testbox-lease-freshness.mjs
index 22e1da91ba8f..dfa2a1c6c968 100644
--- a/scripts/testbox-lease-freshness.mjs
+++ b/scripts/testbox-lease-freshness.mjs
@@ -86,12 +86,6 @@ export function buildTestboxLeaseFingerprint(repoRoot, args) {
   };
 }
 
-export function testboxLeaseCanSkipSync(saved, current) {
-  return Boolean(
-    saved?.syncedCleanHead && current.workingTreeClean && saved.syncedCleanHead === current.headSha,
-  );
-}
-
 export function testboxLeaseStaleReasons(saved, current) {
   if (!saved || saved.version !== STATE_VERSION) {
     return ["state schema"];
@@ -121,9 +115,9 @@ export function prepareTestboxLeaseFreshness({ args, env, provider, repoRoot })
         `Testbox ${id} is stale (${staleReasons.join(", ")}); stop it and warm a fresh lease, or set OPENCLAW_TESTBOX_ALLOW_STALE=1 for an intentional diagnostic reuse`,
       );
     }
-    return { current, path, skipSync: testboxLeaseCanSkipSync(saved, current) };
+    return { current, path };
   }
-  return { current, path, skipSync: false };
+  return { current, path };
 }
 
 export function recordTestboxLeaseFreshness(prepared) {
@@ -132,10 +126,6 @@ export function recordTestboxLeaseFreshness(prepared) {
   }
   mkdirSync(resolve(prepared.path, ".."), { recursive: true });
   const temporaryPath = `${prepared.path}.tmp-${process.pid}`;
-  const state = {
-    ...prepared.current,
-    syncedCleanHead: prepared.current.workingTreeClean ? prepared.current.headSha : "",
-  };
-  writeFileSync(temporaryPath, `${JSON.stringify(state, null, 2)}\n`);
+  writeFileSync(temporaryPath, `${JSON.stringify(prepared.current, null, 2)}\n`);
   renameSync(temporaryPath, prepared.path);
 }
diff --git a/test/scripts/testbox-lease-freshness.test.ts b/test/scripts/testbox-lease-freshness.test.ts
index fe5cc9cc107e..12574b671cef 100644
--- a/test/scripts/testbox-lease-freshness.test.ts
+++ b/test/scripts/testbox-lease-freshness.test.ts
@@ -1,8 +1,5 @@
 import { describe, expect, it } from "vitest";
-import {
-  testboxLeaseCanSkipSync,
-  testboxLeaseStaleReasons,
-} from "../../scripts/testbox-lease-freshness.mjs";
+import { testboxLeaseStaleReasons } from "../../scripts/testbox-lease-freshness.mjs";
 
 const fingerprint = {
   version: 1,
@@ -21,24 +18,6 @@ describe("Testbox lease freshness", () => {
     expect(testboxLeaseStaleReasons(fingerprint, { ...fingerprint })).toEqual([]);
   });
 
-  it("skips sync only for the exact clean HEAD already proven on the lease", () => {
-    expect(
-      testboxLeaseCanSkipSync(
-        { ...fingerprint, syncedCleanHead: fingerprint.headSha },
-        fingerprint,
-      ),
-    ).toBe(true);
-    expect(
-      testboxLeaseCanSkipSync(
-        { ...fingerprint, syncedCleanHead: fingerprint.headSha },
-        { ...fingerprint, workingTreeClean: false },
-      ),
-    ).toBe(false);
-    expect(
-      testboxLeaseCanSkipSync({ ...fingerprint, syncedCleanHead: "e".repeat(40) }, fingerprint),
-    ).toBe(false);
-  });
-
   it("rotates a lease when base, dependency, or workflow inputs drift", () => {
     expect(
       testboxLeaseStaleReasons(fingerprint, {