From 17449fa5197a47e6d70137a6dc3ee11f7ab719d5 Mon Sep 17 00:00:00 2001
From: Eve <eve@ai.cowbay.org>
Date: Thu, 23 Apr 2026 14:49:15 +0800
Subject: [PATCH] feat: block progress claims without execution evidence

---
 hooks/force-recall/handler.ts                     |  6 ++++++
 scripts/long_task_gate_lock.mjs                   | 13 ++++++++++++-
 scripts/test_force_recall_long_task_preflight.mjs |  2 ++
 3 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/hooks/force-recall/handler.ts b/hooks/force-recall/handler.ts
index 8a60d2b..f2e1bd6 100644
--- a/hooks/force-recall/handler.ts
+++ b/hooks/force-recall/handler.ts
@@ -105,6 +105,11 @@ function buildGateLockInput(wrapperResult: any): Record<string, unknown> {
       "bind_externalized_checkpoint_path_or_abort_silent_launch",
     ].includes(requiredNextAction),
   );
+  const executionEvidence = hasConcreteExecutionEvidence
+    ? {
+        concreteNextAction: requiredNextAction,
+      }
+    : null;
   const hasExternalizedCheckpointEvidence = wrapperResult.silentLaunchOk === true
     && typeof wrapperResult.taskRecord?.task_name === "string"
     && wrapperResult.taskRecord.task_name.trim().length > 0;
@@ -124,6 +129,7 @@ function buildGateLockInput(wrapperResult: any): Record<string, unknown> {
     progressionClaim: claimedProgression,
     claimedProgression: claimedProgression,
     statusSummary: claimedProgression,
+    executionEvidence,
     toolCallEvidence: "",
     dispatchEvidence: "",
     fileChangeEvidence: "",
diff --git a/scripts/long_task_gate_lock.mjs b/scripts/long_task_gate_lock.mjs
index ec18672..87edab6 100644
--- a/scripts/long_task_gate_lock.mjs
+++ b/scripts/long_task_gate_lock.mjs
@@ -23,6 +23,7 @@ const EVIDENCE_FIELDS = Object.freeze({
     'statusSummary',
   ]),
   executionEvidence: Object.freeze([
+    'executionEvidence',
     'toolCallEvidence',
     'dispatchEvidence',
     'fileChangeEvidence',
@@ -153,6 +154,16 @@ function usesButtonPath(input) {
   return hasAcceptedValue(input, EVIDENCE_FIELDS.buttonPathMode, 'button_path');
 }
 
+function hasExecutionEvidence(input) {
+  return EVIDENCE_FIELDS.executionEvidence.some((fieldPath) => {
+    const value = getPathValue(input, fieldPath);
+    if (hasNonEmptyString(value)) return true;
+    if (Array.isArray(value)) return value.length > 0;
+    if (value && typeof value === 'object') return Object.keys(value).length > 0;
+    return false;
+  });
+}
+
 function claimsProgressionWithoutEvidence(input) {
   const progressionClaim = EVIDENCE_FIELDS.progressionClaim
     .map((fieldPath) => getPathValue(input, fieldPath))
@@ -160,7 +171,7 @@ function claimsProgressionWithoutEvidence(input) {
 
   if (!hasNonEmptyString(progressionClaim)) return false;
 
-  return !hasAnyNonEmptyString(input, EVIDENCE_FIELDS.executionEvidence);
+  return !hasExecutionEvidence(input);
 }
 
 function evaluateGate(input) {
diff --git a/scripts/test_force_recall_long_task_preflight.mjs b/scripts/test_force_recall_long_task_preflight.mjs
index a0476d9..e36aead 100644
--- a/scripts/test_force_recall_long_task_preflight.mjs
+++ b/scripts/test_force_recall_long_task_preflight.mjs
@@ -66,6 +66,8 @@ async function main() {
     'reason=silent long-task cannot continue without externalized checkpoint path',
     'reason=claimed execution requires evidence of a concrete next action',
     'reason=owner decision flow must end in button-path, not plain text',
+    'reason=claimed progression without concrete execution evidence is forbidden',
+    'requiredEvidence=executionEvidence',
     'HARD_GATE: Block any plain-text handoff or silent-continuation claim when externalized checkpoint evidence is missing.',
     'HARD_GATE: Block any reply path that says you already moved into the next task or are advancing the next step without concrete execution evidence.',
     'HARD_GATE: Do not say you are already on the next task, already dispatched follow-up work, or already progressing in background unless you can point to actual tool execution, file changes, emitted messages, or checkpoint records.',