From eccf61fbc934d62fe78e2ee6bc0bdf6d8a7d7421 Mon Sep 17 00:00:00 2001
From: OG T <ogt@WOOOMacMiniM4.local>
Date: Wed, 1 Apr 2026 15:59:42 +0800
Subject: [PATCH] =?UTF-8?q?fix(ai):=20=E4=BF=AE=E5=BE=A9=E5=81=87=E4=BF=A1?=
 =?UTF-8?q?=E5=BF=83=E5=BA=A6=20+=20=E8=A7=A3=E9=99=A4=20Shadow=20Mode=20(?=
 =?UTF-8?q?Phase=2022=20P1)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

1. openclaw.py: LLM 截斷時 confidence 0.82→0.0 (禁止偽造信心度)
2. prompts.py: NEMOTRON schema 範例值改用佔位符，防模型照抄 0.75
3. configmap: SHADOW_MODE_ENABLED=false，開放 low 風險自動執行
   條件門檻: confidence≥90% + trust_score≥5 + playbook_success≥95%

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---
 apps/api/src/core/prompts.py      | 4 ++--
 apps/api/src/services/openclaw.py | 5 +++--
 k8s/awoooi-prod/04-configmap.yaml | 7 +++++++
 3 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/apps/api/src/core/prompts.py b/apps/api/src/core/prompts.py
index 44a1e2e2..ea9d7a82 100644
--- a/apps/api/src/core/prompts.py
+++ b/apps/api/src/core/prompts.py
@@ -140,7 +140,7 @@ You are an SRE AI. Analyze the alert and respond with ONLY valid JSON.
 
 ## Required JSON Schema:
 {
-  "confidence": 0.75,
+  "confidence": <YOUR_CALCULATED_VALUE>,
   "reasoning": "簡短理由 (繁體中文)",
   "primary_responsibility": "FE|BE|INFRA|DB|COLLAB",
   "risk_level": "low|medium|critical",
@@ -155,7 +155,7 @@ You are an SRE AI. Analyze the alert and respond with ONLY valid JSON.
 
 ## Rules:
 1. Response MUST be valid JSON.
-2. confidence MUST be a real number you calculated (0.0-1.0). NOT 0.75 or 0.82.
+2. confidence is a float 0.0-1.0 you CALCULATE from evidence. High evidence = 0.85-0.95. Low evidence = 0.40-0.65. NEVER copy example values.
 3. Language: Traditional Chinese (Taiwan).
 4. No explanation outside JSON.
 """
diff --git a/apps/api/src/services/openclaw.py b/apps/api/src/services/openclaw.py
index 997b2755..e3d752e6 100644
--- a/apps/api/src/services/openclaw.py
+++ b/apps/api/src/services/openclaw.py
@@ -1041,9 +1041,10 @@ class OpenClawService:
             if "suggested_action" not in data:
                 data["suggested_action"] = "NO_ACTION"
 
-            # Step 2.5: 2026-03-31 ogt - 強力補全計畫 (針對 Nemo-4B 斷片)
+            # Step 2.5: 2026-04-01 Claude Code - 斷片補全 (信心度必須誠實)
+            # 🔴 禁止填入假信心度！截斷 = 0.0，讓 auto-approve 正確判斷
             if "confidence" not in data or not isinstance(data["confidence"], int | float):
-                data["confidence"] = 0.82  # 給予合理平均值而非 0
+                data["confidence"] = 0.0  # 截斷/缺失 → 0.0，不可偽造
             if "risk_level" not in data:
                 data["risk_level"] = "low"
             if "primary_responsibility" not in data:
diff --git a/k8s/awoooi-prod/04-configmap.yaml b/k8s/awoooi-prod/04-configmap.yaml
index 417b13d0..d9a15c0d 100644
--- a/k8s/awoooi-prod/04-configmap.yaml
+++ b/k8s/awoooi-prod/04-configmap.yaml
@@ -48,6 +48,13 @@ data:
   ENABLE_NEMOTRON_COLLABORATION: "true"
   NEMOTRON_TIMEOUT_SECONDS: "45"
   NEMOTRON_ASYNC_UPDATE: "true"
+
+  # ============================================================================
+  # Phase 22 P1: Shadow Mode 解除 - 開放低風險自動執行 (2026-04-01 ogt)
+  # 條件: risk_level=low AND confidence≥90% AND trust_score≥5 AND playbook_success≥95%
+  # ============================================================================
+  SHADOW_MODE_ENABLED: "false"
+  SHADOW_MODE_LOG_ONLY: "false"
   # K8s 接管 @tsenyangbot Long Polling (停用 OpenClaw 188 Telegram)
   TELEGRAM_ENABLE_POLLING: "true"