scene-capability-engine 3.6.65 → 3.6.67

package/CHANGELOG.md

@@ -7,6 +7,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [3.6.67] - 2026-03-29
+
+### Fixed
+- Relaxed `audit:agent-governance` so a clean-room checkout no longer hard-fails release validation just because the gitignored personal file `.sce/steering/CURRENT_CONTEXT.md` is absent.
+
+## [3.6.66] - 2026-03-28
+
+### Added
+- Added `audit:deprecated-entry` plus `report:release-posture`/`gate:release-posture` so top-level entry docs and stable-release posture are now machine-checked instead of being left to manual interpretation.
+- Added a dedicated `release-extended-evidence.yml` workflow so Moqui, matrix, weekly-ops, and other heavy release evidence can be generated independently from the stable publish gate.
+
+### Changed
+- Rebuilt `release.yml` into a stable-core release pipeline that gates publish on release tests, governance audits, deprecated-entry audit, git-managed gate, errorbook gates, and stable posture verification.
+- Rewrote top-level integration entry docs and template entry docs so SCE now leads with the governed `project -> channel/session -> scene/spec` model, while `context export / prompt generate` stay documented only as compatibility paths.
+- Promoted current release posture from `rc-candidate` to `stable` after separating stable-core and extended evidence, and after wiring deprecated-entry governance into CI/release baselines.
+
 ## [3.6.65] - 2026-03-22
 
 ### Added

package/README.md

@@ -16,7 +16,7 @@ AI agents can generate code quickly, but they also drift, over-create context, l
 
 SCE provides the missing control layer:
 
-- `Scene-governed context`: one primary session per scene, many specs per scene, many tasks per spec.
+- `Scene-governed context`: `scene` stays the semantic boundary, while one project can host many channels/sessions in parallel and each channel keeps its own scene + spec focus.
 - `Spec-first execution`: requirements, design, tasks, and gates stay attached to the work, not buried in chat history.
 - `Bounded autonomous delivery`: agents can run `close-loop`, `close-loop-program`, and `close-loop-controller` with retry, fallback, and governance policies.
 - `Recoverable local history`: timeline snapshots, task refs, and SQLite-backed state keep work recoverable even before Git push.
@@ -26,10 +26,12 @@ SCE provides the missing control layer:
 
 ## Core Model
 
-SCE organizes agent work using one stable hierarchy:
+SCE organizes agent work using one stable semantic/collaboration model:
 
-- `session -> scene -> spec -> task -> event`
+- `project -> scene -> spec`
+- `project -> channel/session -> (scene + spec focus) -> task -> event`
 - `scene` is the continuity boundary
+- `channel/session` is the collaboration thread, not the semantic domain
 - `spec` is the governed work package
 - `task` is the smallest user-facing execution unit
 - `event` remains the raw audit stream behind the task view
@@ -49,7 +51,7 @@ This gives you a predictable way to manage long-running agent work without relyi
 ### 2. Studio Execution Flow
 - `studio plan -> generate -> apply -> verify -> release`
 - Structured task stream for frontend or IDE integration
-- Task refs (`SS.PP.TT`) for lookup and rerun
+- Hierarchical task refs (`scene.spec.task` / `SS.PP.TT`) for lookup, controlled rerun, and rerun hints aligned to the original `project + channel` focus
 - Auth lease model for protected write operations
 
 ### 3. Autonomous Delivery
@@ -74,6 +76,11 @@ This gives you a predictable way to manage long-running agent work without relyi
 ### 6. Capability and Scene Assetization
 - Scene/capability inventory and governance views
 - Capability extraction, evaluation, and publication workflow
+- Semantic observation, replay evaluation, local promotion, governed delta export, and native backflow rerank
+- Shared-library mirror publication that advances qualified lessons from `publish_pending` to `published-shared`
+- `semantic promote` now auto-publishes qualified lessons by default under the adopt-implied model
+- Shared-sync handoff bundle export that packages the local semantic mirror into spec-scoped bundle/index/shards/receipt artifacts
+- Shared-sync intake that lets another terminal sync governed bundle output back into its local semantic mirror
 - Scene runtime and ontology-oriented execution support
 - Moqui-oriented capability validation and handoff baselines
 
@@ -139,7 +146,11 @@ sce auth status --json
 SCE is opinionated by default.
 
 - `studio plan` runs intake and scene/spec governance unless policy explicitly allows bypass.
+- `studio plan`, `studio intake`, and `native ask` pass through the Four Teachings supreme-intent gate first; off-track requests are clarified, rewritten, narrowed, or refused before execution, and existing bound Specs receive governed evidence.
 - When business scene/module/page/entity context is missing, SCE must route to clarification first; unknown business scope must not be turned into blanket disable.
+- `CORE_PRINCIPLES.md` keeps only `Four Teachings + Little Nine` in the supreme-principle layer; execution baselines belong in `RULES_GUIDE.md`, and reusable tactics belong in governance docs or lessons.
+- After two failed localization rounds, record/update an `errorbook` incident and use bisection-style debug logs to keep halving the failing scope instead of continuing blind patch attempts.
+- Agent-generated scripts, reports, debug logs, test scripts, and temporary analysis default to the active Spec subdirectories; if no explicit Spec exists yet, use a governed general Spec instead of scattering files.
 - `spec pipeline` and `spec gate` now carry a read-only strategy advisory for single-Spec runs, so broad or clarification-heavy work is surfaced as `multi-spec-program` / `research-program` instead of being blindly forced deeper into one Spec.
 - `verify` and `release` enforce problem-closure and related gates when a spec is bound.
 - Autonomous program execution applies gate evaluation, fallback-chain logic, governance replay, and auto-remediation.
@@ -230,5 +241,5 @@ MIT. See [LICENSE](LICENSE).
 
 ---
 
-**Version**: 3.6.65
-**Last Updated**: 2026-03-22
+**Version**: 3.6.67
+**Last Updated**: 2026-03-29

package/README.zh.md

@@ -21,7 +21,7 @@ AI Agent 可以很快生成代码，但也很容易出现这些问题：
 
 SCE 提供的就是这一层控制能力：
 
-- `Scene 主导上下文`：一个 scene 对应一个主会话，多个 spec 挂在 scene 下持续推进
+- `Scene 主导上下文`：`scene` 是稳定语义边界，一个项目可并行存在多条 `channel/session`，每条通道各自维护当前 `scene + spec` 焦点
 - `Spec 先行治理`：需求、设计、任务、门禁都绑定到 spec，而不是散落在聊天记录里
 - `自动闭环可控`：`close-loop`、`close-loop-program`、`close-loop-controller` 都能在边界内自动收敛
 - `本地历史可恢复`：timeline、task ref、SQLite 状态让过程资产可回看、可恢复、可重跑
@@ -31,10 +31,12 @@ SCE 提供的就是这一层控制能力：
 
 ## 核心对象模型
 
-SCE 用一条稳定层级来管理 Agent 工作：
+SCE 用一条稳定的“语义链 + 协作链”来管理 Agent 工作：
 
-- `session -> scene -> spec -> task -> event`
+- `project -> scene -> spec`
+- `project -> channel/session -> (scene + spec focus) -> task -> event`
 - `scene` 是业务连续性的主边界
+- `channel/session` 是协作线程，不是问题域本身
 - `spec` 是被治理的工作包
 - `task` 是面向人的最小执行单元
 - `event` 保留原始审计流，作为 task 背后的底层记录
@@ -54,7 +56,7 @@ SCE 用一条稳定层级来管理 Agent 工作：
 ### 2. Studio 执行流
 - `studio plan -> generate -> apply -> verify -> release`
 - 面向 IDE / 前端的结构化 task 流
-- 任务引用 `SS.PP.TT`，支持查询和重跑
+- 层级任务引用（`scene.spec.task` / `SS.PP.TT`），支持查询、受控重跑与按原 `project + channel` 焦点生成重跑提示
 - 写操作支持 auth lease 保护
 
 ### 3. 自动闭环交付
@@ -79,6 +81,11 @@ SCE 用一条稳定层级来管理 Agent 工作：
 ### 6. 能力资产化与场景化能力
 - scene/capability inventory 与治理视图
 - 能力抽取、评估、发布闭环
+- 支持语义观测、replay 评估、本地推广、共享 delta 导出与 native 回流重排
+- 支持 shared-library mirror 发布，把合格能力从 `publish_pending` 推进到 `published-shared`
+- `semantic promote` 默认按 adopt-implied 模型立即自动共享发布合格能力
+- 支持 shared-sync handoff bundle 导出，把本地 semantic mirror 打包成 spec 内受治理的 bundle/index/shards/receipt
+- 支持 shared-sync intake，把其他终端导出的受治理 bundle 同步回本地 semantic mirror
 - 支持 scene runtime 与 ontology 驱动执行
 - 支持面向 Moqui 的能力校验与 handoff 基线
 
@@ -145,11 +152,16 @@ SCE 默认是强治理的。
 
 - `studio plan` 默认执行 intake 与 scene/spec 治理，除非策略显式允许绕过
 - 缺少业务场景/模块/页面/实体上下文时，SCE 必须先进入澄清，而不是把未知业务范围直接变成一刀切禁用
+- 连续两轮以上仍未成功定位问题时，SCE 默认要求先补 `errorbook` incident，再用二分式 debug 日志快速收敛范围
 - 单 spec 的 `spec pipeline` 和 `spec gate` 现在都会附带只读策略建议；当问题已经更适合 `multi-spec-program` / `research-program` 时，SCE 会显式提示，而不是继续把复杂问题盲目压进同一个 spec
 - 当 spec 绑定时，`verify` 和 `release` 默认执行 problem-closure 等相关门禁
 - `close-loop-program` 默认带 gate 评估、fallback-chain、governance replay、auto-remediation
 - co-work 基线默认开启：初始化或接管后的 SCE 项目会落地 `.sce/config/multi-agent.json` 且 `enabled=true`，但中央 coordinator 仍保持按需开启
 - 状态持久化默认优先走 SQLite，而不是零散本地缓存
+- `CORE_PRINCIPLES.md` 的最高原则层默认只保留“四句教 + 小九条”；四句教先判断该不该做，小九条再判断如何把正确的事做对；治理基线进入 `RULES_GUIDE.md`，执行经验与案例进入 `docs/steering-governance.md` 或 `.sce/knowledge/lessons/`
+- `studio plan` / `studio intake` / `native ask` 默认先经过“四句教”判定闸门；对失范目标会执行 `clarify / rewrite / narrow / refuse`，并把判定结果写入 session/job/spec 证据链
+- Agent 生成的脚本、报告、调试日志、测试脚本和临时分析默认进入当前 Spec 子目录；没有明确 Spec 时先用通用 Spec 承接
+- 默认遵守“小九条”执行姿态：勤奋推进、先领会真实意图、先看事实、维护项目活地图、想透问题、证据足够时敢下判断、统筹 Spec/错题本/测试/现状代码/关注点、先拆解定位阻塞、持续收敛直到闭环
 - 超大源文件必须定期触发重构评估；SCE 优先建议按项目给出阈值，若项目尚未设定，则默认参考 `2000 / 4000 / 10000`
 - 发布默认验证走 integration gate：`npm run test:release`
 
@@ -235,5 +247,5 @@ MIT，见 [LICENSE](LICENSE)。
 
 ---
 
-**版本**：3.6.65
-**最后更新**：2026-03-22
+**版本**：3.6.67
+**最后更新**：2026-03-29

package/bin/scene-capability-engine.js

@@ -28,6 +28,8 @@ const { registerTaskCommands } = require('../lib/commands/task');
 const { registerAuthCommands } = require('../lib/commands/auth');
 const { registerStateCommands } = require('../lib/commands/state');
 const { registerCapabilityCommands } = require('../lib/commands/capability');
+const { registerSemanticCommands } = require('../lib/commands/semantic');
+const { registerNativeCommands } = require('../lib/commands/native');
 const VersionChecker = require('../lib/version/version-checker');
 const {
   findLegacyKiroDirectories,
@@ -1062,6 +1064,8 @@ registerTaskCommands(program);
 registerAuthCommands(program);
 registerStateCommands(program);
 registerCapabilityCommands(program);
+registerSemanticCommands(program);
+registerNativeCommands(program);
 
 const { registerAppCommands } = require('../lib/commands/app');
 const { registerDeviceCommands } = require('../lib/commands/device');

package/docs/README.md

@@ -72,7 +72,7 @@ This hub keeps only the high-value entry points for adopting, integrating, and s
 
 ---
 
-**Version**: 3.6.65
-**Last Updated**: 2026-03-22
+**Version**: 3.6.67
+**Last Updated**: 2026-03-29