Integration-Automation · JE-Chen · Jun 21, 2026 · Jun 21, 2026
diff --git a/README.md b/README.md
@@ -13,6 +13,7 @@
 
 ## Table of Contents
 
+- [What's new (2026-06-21) — Feature Flags](#whats-new-2026-06-21--feature-flags)
 - [What's new (2026-06-21) — Text Diff, Patch & Three-Way Merge](#whats-new-2026-06-21--text-diff-patch--three-way-merge)
 - [What's new (2026-06-21) — Calendar Recurrence Rules (RRULE)](#whats-new-2026-06-21--calendar-recurrence-rules-rrule)
 - [What's new (2026-06-21) — Statistics & A/B Significance](#whats-new-2026-06-21--statistics--ab-significance)
@@ -120,6 +121,12 @@
 
 ---
 
+## What's new (2026-06-21) — Feature Flags
+
+Toggle behavior with targeting & rollout. Full reference: [`docs/source/Eng/doc/new_features/v68_features_doc.rst`](docs/source/Eng/doc/new_features/v68_features_doc.rst).
+
+- **`FlagStore` / `evaluate_flag` / `is_enabled` / `assign_variant`** (`AC_evaluate_flag`, `AC_flag_enabled`): `decision_table` is one-shot DMN and `ab_locator` is locator A/B — neither is a product flag store with sticky % rollout. This adds an OpenFeature-shaped engine: targeting rules (`eq`/`in`/`semver_*`…), weighted variants, kill switch, and consistent-hash bucketing (`sha256(key.salt.context_key)`) so a subject is **sticky**. Returns `{value, variant, reason}` (`TARGETING_MATCH`/`SPLIT`/`DISABLED`/`ERROR`). Pure-stdlib, deterministic.
+
 ## What's new (2026-06-21) — Text Diff, Patch & Three-Way Merge
 
 Apply and merge text diffs. Full reference: [`docs/source/Eng/doc/new_features/v67_features_doc.rst`](docs/source/Eng/doc/new_features/v67_features_doc.rst).

diff --git a/README/README_zh-CN.md b/README/README_zh-CN.md
@@ -12,6 +12,7 @@
 
 ## 目录
 
+- [本次更新 (2026-06-21) — 功能旗标](#本次更新-2026-06-21--功能旗标)
 - [本次更新 (2026-06-21) — 文本 Diff、应用与三方合并](#本次更新-2026-06-21--文本-diff应用与三方合并)
 - [本次更新 (2026-06-21) — 日历周期规则(RRULE)](#本次更新-2026-06-21--日历周期规则rrule)
 - [本次更新 (2026-06-21) — 统计与 A/B 显著性](#本次更新-2026-06-21--统计与-ab-显著性)
@@ -119,6 +120,12 @@
 
 ---
 
+## 本次更新 (2026-06-21) — 功能旗标
+
+以目标规则与推出切换行为。完整参考:[`docs/source/Zh/doc/new_features/v68_features_doc.rst`](../docs/source/Zh/doc/new_features/v68_features_doc.rst)。
+
+- **`FlagStore` / `evaluate_flag` / `is_enabled` / `assign_variant`**(`AC_evaluate_flag`、`AC_flag_enabled`):`decision_table` 是一次性 DMN,`ab_locator` 是定位器 A/B —— 两者都不是带黏性 % 推出的产品旗标存储库。本功能补上 OpenFeature 形状引擎:目标规则(`eq`/`in`/`semver_*`…)、加权变体、kill switch,以及一致哈希分桶(`sha256(key.salt.context_key)`)使主体具**黏性**。返回 `{value, variant, reason}`(`TARGETING_MATCH`/`SPLIT`/`DISABLED`/`ERROR`)。纯标准库、确定。
+
 ## 本次更新 (2026-06-21) — 文本 Diff、应用与三方合并
 
 应用并合并文本 diff。完整参考:[`docs/source/Zh/doc/new_features/v67_features_doc.rst`](../docs/source/Zh/doc/new_features/v67_features_doc.rst)。

diff --git a/README/README_zh-TW.md b/README/README_zh-TW.md
@@ -12,6 +12,7 @@
 
 ## 目錄
 
+- [本次更新 (2026-06-21) — 功能旗標](#本次更新-2026-06-21--功能旗標)
 - [本次更新 (2026-06-21) — 文字 Diff、套用與三方合併](#本次更新-2026-06-21--文字-diff套用與三方合併)
 - [本次更新 (2026-06-21) — 行事曆週期規則(RRULE)](#本次更新-2026-06-21--行事曆週期規則rrule)
 - [本次更新 (2026-06-21) — 統計與 A/B 顯著性](#本次更新-2026-06-21--統計與-ab-顯著性)
@@ -119,6 +120,12 @@
 
 ---
 
+## 本次更新 (2026-06-21) — 功能旗標
+
+以目標規則與推出切換行為。完整參考:[`docs/source/Zh/doc/new_features/v68_features_doc.rst`](../docs/source/Zh/doc/new_features/v68_features_doc.rst)。
+
+- **`FlagStore` / `evaluate_flag` / `is_enabled` / `assign_variant`**(`AC_evaluate_flag`、`AC_flag_enabled`):`decision_table` 是一次性 DMN,`ab_locator` 是定位器 A/B —— 兩者都不是帶黏性 % 推出的產品旗標儲存庫。本功能補上 OpenFeature 形狀引擎:目標規則(`eq`/`in`/`semver_*`…)、加權變體、kill switch,以及一致雜湊分桶(`sha256(key.salt.context_key)`)使主體具**黏性**。回傳 `{value, variant, reason}`(`TARGETING_MATCH`/`SPLIT`/`DISABLED`/`ERROR`)。純標準函式庫、具決定性。
+
 ## 本次更新 (2026-06-21) — 文字 Diff、套用與三方合併
 
 套用並合併文字 diff。完整參考:[`docs/source/Zh/doc/new_features/v67_features_doc.rst`](../docs/source/Zh/doc/new_features/v67_features_doc.rst)。

diff --git a/docs/source/Eng/doc/new_features/v68_features_doc.rst b/docs/source/Eng/doc/new_features/v68_features_doc.rst
@@ -0,0 +1,58 @@
+Feature Flags
+=============
+
+``decision_table`` is a one-shot DMN evaluator and ``ab_locator`` measures
+locator outcomes — neither is a product feature-flag store with sticky
+percentage rollout. This adds an OpenFeature-shaped flag engine: typed flags
+with targeting rules, weighted variants, a kill switch, and consistent-hash
+bucketing so a given subject always lands in the same variant.
+
+Pure standard library (``hashlib`` + ``re`` + ``json``); deterministic; imports
+no ``PySide6``.
+
+Headless API
+------------
+
+.. code-block:: python
+
+    from je_auto_control import FlagStore, evaluate_flag, is_enabled
+
+    store = FlagStore.from_dict({"flags": {
+        "new-checkout": {
+            "variants": {"on": True, "off": False},
+            "default_variant": "off", "off_variant": "off",
+            "targeting": [
+                {"conditions": {"country": {"op": "in", "value": ["US", "CA"]}},
+                 "serve": "on"},
+                {"conditions": {"plan": {"op": "eq", "value": "premium"}},
+                 "serve": {"rollout": {"on": 50, "off": 50}}},
+            ],
+            "fallthrough": {"rollout": {"on": 10, "off": 90}},
+        },
+    }})
+
+    evaluate_flag(store, "new-checkout", {"country": "US"})
+    # {"flag_key": "new-checkout", "variant": "on", "value": True,
+    #  "reason": "TARGETING_MATCH"}
+
+    if is_enabled(store, "new-checkout", {"targeting_key": "user-123"}):
+        ...
+
+Evaluation order mirrors OpenFeature/Unleash/LaunchDarkly: a disabled flag
+serves ``off_variant`` (reason ``DISABLED``); an unknown flag returns the
+caller default (reason ``ERROR``); targeting rules are tried in order
+(``TARGETING_MATCH``); otherwise the fallthrough applies (``DEFAULT`` /
+``SPLIT``). Targeting operators include ``eq``/``ne``/``lt``/``gt``/``in``/
+``not_in``/``contains`` and ``semver_*``. Percentage rollout is a
+consistent-hash bucket of ``sha256("{key}.{salt}.{context_key}")`` so a subject
+is **sticky** — it always gets the same variant. ``percentage_bucket`` and
+``assign_variant`` are exposed for direct use.
+
+Executor commands
+-----------------
+
+``AC_evaluate_flag`` takes ``flags`` (a store mapping or JSON string), a
+``key`` and optional ``context``; it returns ``{value, variant, reason}``.
+``AC_flag_enabled`` returns ``{enabled}``. Both are exposed as MCP tools
+(``ac_evaluate_flag`` / ``ac_flag_enabled``) and as Script Builder commands
+under **Flow**.
diff --git a/docs/source/Eng/eng_index.rst b/docs/source/Eng/eng_index.rst
@@ -90,6 +90,7 @@ Comprehensive guides for all AutoControl features.
    doc/new_features/v65_features_doc
    doc/new_features/v66_features_doc
    doc/new_features/v67_features_doc
+   doc/new_features/v68_features_doc
    doc/ocr_backends/ocr_backends_doc
    doc/observability/observability_doc
    doc/operations_layer/operations_layer_doc

diff --git a/docs/source/Zh/doc/new_features/v68_features_doc.rst b/docs/source/Zh/doc/new_features/v68_features_doc.rst
@@ -0,0 +1,50 @@
+功能旗標(Feature Flags)
+========================
+
+``decision_table`` 是一次性的 DMN 評估器,``ab_locator`` 衡量定位器成效 —— 兩者都不是帶有黏性
+百分比推出的產品功能旗標儲存庫。本功能補上 OpenFeature 形狀的旗標引擎:具型別的旗標、目標規則、
+加權變體、kill switch,以及一致雜湊分桶,讓同一主體永遠落在同一變體。
+
+純標準函式庫(``hashlib`` + ``re`` + ``json``);具決定性;不匯入 ``PySide6``。
+
+無頭 API
+--------
+
+.. code-block:: python
+
+    from je_auto_control import FlagStore, evaluate_flag, is_enabled
+
+    store = FlagStore.from_dict({"flags": {
+        "new-checkout": {
+            "variants": {"on": True, "off": False},
+            "default_variant": "off", "off_variant": "off",
+            "targeting": [
+                {"conditions": {"country": {"op": "in", "value": ["US", "CA"]}},
+                 "serve": "on"},
+                {"conditions": {"plan": {"op": "eq", "value": "premium"}},
+                 "serve": {"rollout": {"on": 50, "off": 50}}},
+            ],
+            "fallthrough": {"rollout": {"on": 10, "off": 90}},
+        },
+    }})
+
+    evaluate_flag(store, "new-checkout", {"country": "US"})
+    # {"flag_key": "new-checkout", "variant": "on", "value": True,
+    #  "reason": "TARGETING_MATCH"}
+
+    if is_enabled(store, "new-checkout", {"targeting_key": "user-123"}):
+        ...
+
+評估順序對齊 OpenFeature/Unleash/LaunchDarkly:停用的旗標提供 ``off_variant``(原因
+``DISABLED``);未知旗標回傳呼叫端預設(原因 ``ERROR``);目標規則依序嘗試(``TARGETING_MATCH``);
+否則套用 fallthrough(``DEFAULT`` / ``SPLIT``)。目標運算子包含 ``eq``/``ne``/``lt``/``gt``/
+``in``/``not_in``/``contains`` 與 ``semver_*``。百分比推出是
+``sha256("{key}.{salt}.{context_key}")`` 的一致雜湊分桶,因此主體具**黏性** —— 永遠得到相同變體。
+``percentage_bucket`` 與 ``assign_variant`` 亦可直接使用。
+
+執行器命令
+----------
+
+``AC_evaluate_flag`` 接受 ``flags``(store 對映或 JSON 字串)、``key`` 與選用的 ``context``,回傳
+``{value, variant, reason}``。``AC_flag_enabled`` 回傳 ``{enabled}``。兩者皆以 MCP 工具
+(``ac_evaluate_flag`` / ``ac_flag_enabled``)以及 Script Builder 中 **Flow** 分類下的命令提供。
diff --git a/docs/source/Zh/zh_index.rst b/docs/source/Zh/zh_index.rst
@@ -90,6 +90,7 @@ AutoControl 所有功能的完整使用指南。
    doc/new_features/v65_features_doc
    doc/new_features/v66_features_doc
    doc/new_features/v67_features_doc
+   doc/new_features/v68_features_doc
    doc/ocr_backends/ocr_backends_doc
    doc/observability/observability_doc
    doc/operations_layer/operations_layer_doc

diff --git a/je_auto_control/__init__.py b/je_auto_control/__init__.py
@@ -346,6 +346,11 @@
 from je_auto_control.utils.text_diff import (
     MergeResult, PatchApplyError, apply_unified, three_way_merge, unified_diff,
 )
+# Feature flags with targeting rules + deterministic rollout
+from je_auto_control.utils.feature_flags import (
+    Flag, FlagStore, assign_variant, evaluate_flag, is_enabled,
+    percentage_bucket,
+)
 # Background popup/interrupt watchdog (unattended automation)
 from je_auto_control.utils.watchdog import (
     PopupWatchdog, WatchdogRule, default_popup_watchdog,
@@ -834,6 +839,8 @@ def start_autocontrol_gui(*args, **kwargs):
     "Recurrence", "next_occurrence", "occurrences", "parse_rrule",
     "MergeResult", "PatchApplyError", "apply_unified", "three_way_merge",
     "unified_diff",
+    "Flag", "FlagStore", "assign_variant", "evaluate_flag", "is_enabled",
+    "percentage_bucket",
     # MCP server
     "AuditLogger", "HttpMCPServer", "MCPContent", "MCPPrompt",
     "MCPPromptArgument", "MCPResource", "MCPServer", "MCPTool",

diff --git a/je_auto_control/gui/script_builder/command_schema.py b/je_auto_control/gui/script_builder/command_schema.py
@@ -1155,6 +1155,26 @@ def _add_misc_specs(specs: List[CommandSpec]) -> None:
         ),
         description="Validate JSON against a JSON Schema; returns {ok, errors}.",
     ))
+    specs.append(CommandSpec(
+        "AC_evaluate_flag", "Flow", "Feature Flag: Evaluate",
+        fields=(
+            FieldSpec("flags", FieldType.STRING,
+                      placeholder='{"flags": {"f": {"variants": {...}}}}'),
+            FieldSpec("key", FieldType.STRING, placeholder="new-checkout"),
+            FieldSpec("context", FieldType.STRING, optional=True,
+                      placeholder='{"targeting_key": "user1", "country": "US"}'),
+        ),
+        description="Evaluate a feature flag; returns {value, variant, reason}.",
+    ))
+    specs.append(CommandSpec(
+        "AC_flag_enabled", "Flow", "Feature Flag: Enabled?",
+        fields=(
+            FieldSpec("flags", FieldType.STRING),
+            FieldSpec("key", FieldType.STRING, placeholder="new-checkout"),
+            FieldSpec("context", FieldType.STRING, optional=True),
+        ),
+        description="Boolean feature-flag check; returns {enabled}.",
+    ))
     specs.append(CommandSpec(
         "AC_unified_diff", "Data", "Text: Unified Diff",
         fields=(

diff --git a/je_auto_control/utils/executor/action_executor.py b/je_auto_control/utils/executor/action_executor.py
@@ -2928,6 +2928,30 @@ def _rate_limit(name: str, rate: float = 1.0, capacity: float = 1.0,
             "wait": round(bucket.time_until_available(float(n)), 4)}
 
 
+def _evaluate_flag(flags: Any, key: str, context: Any = None) -> Dict[str, Any]:
+    """Adapter: evaluate a feature flag (flags/context dict or JSON string)."""
+    import json
+    from je_auto_control.utils.feature_flags import FlagStore, evaluate_flag
+    if isinstance(flags, str):
+        flags = json.loads(flags)
+    if isinstance(context, str):
+        context = json.loads(context)
+    return evaluate_flag(FlagStore.from_dict(flags), key, context or {})
+
+
+def _flag_enabled(flags: Any, key: str, context: Any = None,
+                  default: bool = False) -> Dict[str, Any]:
+    """Adapter: boolean feature-flag check."""
+    import json
+    from je_auto_control.utils.feature_flags import FlagStore, is_enabled
+    if isinstance(flags, str):
+        flags = json.loads(flags)
+    if isinstance(context, str):
+        context = json.loads(context)
+    store = FlagStore.from_dict(flags)
+    return {"enabled": is_enabled(store, key, context or {}, bool(default))}
+
+
 def _unified_diff(a: str, b: str) -> Dict[str, Any]:
     """Adapter: unified diff transforming text a into b."""
     from je_auto_control.utils.text_diff import unified_diff
@@ -3854,6 +3878,8 @@ def __init__(self):
             "AC_ab_significance": _ab_significance,
             "AC_rrule_occurrences": _rrule_occurrences,
             "AC_rrule_next": _rrule_next,
+            "AC_evaluate_flag": _evaluate_flag,
+            "AC_flag_enabled": _flag_enabled,
             "AC_unified_diff": _unified_diff,
             "AC_apply_unified": _apply_unified,
             "AC_three_way_merge": _three_way_merge,

diff --git a/je_auto_control/utils/feature_flags/__init__.py b/je_auto_control/utils/feature_flags/__init__.py
@@ -0,0 +1,10 @@
+"""Feature-flag evaluation with targeting rules and deterministic rollout."""
+from je_auto_control.utils.feature_flags.feature_flags import (
+    Flag, FlagStore, assign_variant, evaluate_flag, is_enabled,
+    percentage_bucket,
+)
+
+__all__ = [
+    "Flag", "FlagStore", "assign_variant", "evaluate_flag", "is_enabled",
+    "percentage_bucket",
+]