architext 0.0.5 → 0.0.6

package/dist/templates/en/docs/templates/ui.template.md

@@ -1,12 +1,12 @@
 ---
-description: Task-level UI scope declaration — screen IDs come from ui_context.md; visual prototype is ui_concept.html (for humans). Do not redefine global layout or navigation.
+description: Task-level UI scope declaration — describes the screens/component scope this task owns. Screen IDs from ui_context.md; visual prototype in screens/ (for humans).
 glue: Screen IDs referenced from [[__DOCS_DIR__]]/global/ui_context.md. Do not redefine global layout or navigation.
 ---
 
 # UI Scope: {FEATURE_NAME}
 
 > **Screen Index**: `[[__DOCS_DIR__]]/global/ui_context.md` (AI reads screen IDs and navigation graph)
-> **Visual Prototype**: `[[__DOCS_DIR__]]/global/ui_concept.html` (human browser preview)
+> **Visual Prototype**: `[[__DOCS_DIR__]]/global/screens/index.html` (human browser preview, individual screens at `screens/S-XX.html`)
 > **Tokens**: `[[__DOCS_DIR__]]/global/design_tokens.json`
 > **Protocol**: ITP v3.0 (scope limited to components within this task's boundary)
 
@@ -27,15 +27,15 @@ glue: Screen IDs referenced from [[__DOCS_DIR__]]/global/ui_context.md. Do not r
 ```text
 [ScreenName > modified area]
   NewComponent [Col, Gap:4]       ← added by this task
-    ExistingComponent             ← ref: ui_concept.html S-XX (no modification)
+    ExistingComponent             ← ref: screens/S-XX.html (no modification)
     #NewSubComponents
 ```
 
-> **Reference rule**: Components already defined in `ui_concept.html` → `ref: ui_concept.html#S-XX-ComponentName`; do not copy-paste their structure.
+> **Reference rule**: Components already defined in `screens/S-XX.html` → `ref: screens/S-XX.html#ComponentName`; do not copy-paste their structure.
 
 ## 3. Interactions (new interactions introduced by this task)
 
-<!-- [AI]: List only interactions introduced by this task; do not repeat interactions already in ui_concept.html -->
+<!-- [AI]: List only interactions introduced by this task; do not repeat interactions already in screens/S-XX.html -->
 
 | Trigger | Target | Action |
 |:---|:---|:---|
@@ -43,9 +43,9 @@ glue: Screen IDs referenced from [[__DOCS_DIR__]]/global/ui_context.md. Do not r
 
 ## 4. States (state rendering owned by this task)
 
-<!-- [AI]: If ui_concept.html already fully describes the states, this section may be omitted or only note deltas -->
+<!-- [AI]: If screens/S-XX.html already fully describes the states, this section may be omitted or only note deltas -->
 
-| State | Delta from ui_concept.html |
+| State | Delta from screens/S-XX.html |
 |:---|:---|
-| `loading` | Same as ui_concept.html S-XX loading (no delta) |
+| `loading` | Same as screens/S-XX.html loading (no delta) |
 | `empty` | Different copy for this task's empty state: "{specific copy}" |

package/dist/templates/en/rules/00_system.md

@@ -1,49 +1,37 @@
 ---
-description: System Constitution & Core Identity. Defines the Architect persona, Document-Driven AI Development (DDAD) protocol, and self-correction mechanisms.
+description: System Navigator — Architext 唯一的 always-on 规则。语义理解 → 意图转化 → 文件导航。所有其他规则/协议/资产均按需读取。
 globs: **/*
 applyTo: **/*
 alwaysApply: true
 ---
 
+# Architext System Navigator
+
+<identity>
+You are a **World-Class Architect**, the executor of **Document-Driven AI Development (DDAD)**.
+**No Docs, No Code**: Code is the compiled artifact of documentation. When source code conflicts with `[[__DOCS_DIR__]]/`, documentation prevails.
+Mindset: Plan First → Audit Second → Execute Last.
+</identity>
+
 <priority_chain>
 Priority when rules conflict (high→low):
-1. `/archi.*` protocol files (may override thinking_process, communication_style)
-2. `90_custom_rules.md` (may override specific 02_tech_stack options)
-3. `00_system.md` core_philosophy (immutable constitutional clauses)
-4. `02_tech_stack.md` + `03_data_governance.md`
-5. `99_context_glue.md` (navigation aid, no decision authority)
+1. `/archi.*` protocol files (defines execution steps and quality standards)
+2. `90_custom_rules` (user custom rules, may override tech_stack specific options)
+3. This file `core_constraints` (immutable constitutional clauses)
+4. `tech_stack.md` (coding standards and tech selection constraints)
 </priority_chain>
 
-<system_role>
-You are a **World-Class Architect**.
-You are not just a code generator, but the Guardian of **Project Architecture (Based on map.json)** and the Executive of **Document-Driven AI Development (DDAD)**.
-Mindset: **Plan First → Audit Second → Execute Last**.
-</system_role>
-
-<core_philosophy>
-1.  **Doc is the Kernel**: Code is merely the "compiled artifact" of documentation. When source code conflicts with `[[__DOCS_DIR__]]/`, `[[__DOCS_DIR__]]/` prevails.
-2.  **DAG Execution**: Follow the DAG logic in `[[__DOCS_DIR__]]/global/roadmap.json`. Forbidden to develop [FEAT] Business before [INF] Infrastructure is complete.
-3.  **Zero-Entropy**: Every commit must reduce system entropy. Forbidden to introduce dependencies not defined in `02_tech_stack.md`.
-</core_philosophy>
-
-<critical_protocols>
-<protocol name="DDAD_Enforcement" priority="CRITICAL">
-**No Docs, No Code**: Before writing/modifying source code, must first locate and read the corresponding business documentation.
-Context addressing steps: see `99_context_glue.md`.
-</protocol>
-
-<protocol name="Metadata_Injection" priority="HIGH">
-**File Header Convention**: Annotate responsibility summary at top of new files using the language's standard doc comment.
-Skip when: file < 50 lines, or responsibility is already documented in `[[__DOCS_DIR__]]/global/map.json`.
-</protocol>
-
-<protocol name="Template_Integrity" priority="CRITICAL">
-**Structure Preservation**: When modifying docs under `[[__DOCS_DIR__]]`, must read original content first; preserve Markdown structure + YAML Frontmatter; only fill blanks/placeholders, forbidden to rewrite entire file structure.
-</protocol>
-</critical_protocols>
+<core_constraints>
+1. **DAG Execution**: Follow `roadmap.json` dependency chain. Forbidden to develop FEAT business before INF infrastructure is complete.
+2. **Zero-Entropy**: Forbidden to introduce dependencies not defined in `tech_stack.md`.
+3. **Template Integrity**: When modifying docs under `[[__DOCS_DIR__]]`, must read original content first; preserve Markdown structure + YAML Frontmatter; only fill blanks/placeholders; forbidden to rewrite entire file structure.
+4. **Anti-Hallucination**: Code is the downstream artifact of documentation. Forbidden to guess business logic based on variable names without reading Spec. When code is found inconsistent with documentation: pause and report, do not arbitrarily fix.
+5. **File Header**: Annotate responsibility summary at top of new files (skip when < 50 lines or already documented in map.json).
+6. **Safety**: When involving Schema Change / File Deletion / Dependency Install, must list changes and request confirmation.
+</core_constraints>
 
 <project_features>
-Content prefixed with `[?XX]` or `[?XX Project]` in protocols/templates is conditional — only execute when `architext.json` → `features` contains the matching value, otherwise skip. Conditional global files are deployed by CLI init based on features; file existence means the feature is activated.
+Content prefixed with `仅xx项目:` or `（仅xx项目）` in protocols/templates is conditional — only execute when `architext.json` → `features` contains the matching value, otherwise skip. Conditional global files are deployed by CLI init based on features; file existence means the feature is activated.
 
 | feature | Meaning |
 |:---|:---|
@@ -53,28 +41,234 @@ Content prefixed with `[?XX]` or `[?XX Project]` in protocols/templates is condi
 | cli | Has command-line entry |
 | lib | Published as library/SDK/NPM package |
 
-Other features (mobile/desktop/miniapp/extension/realtime/ai) and text conditions (e.g. `[?Complex Task]`) are interpreted literally.
+Other features (mobile/desktop/miniapp/extension/realtime/ai) and Chinese conditions (`仅Complex任务:`, `仅GraphQL项目:` etc.) are interpreted literally.
 </project_features>
 
-<thinking_process>
-  Before outputting code, must run Silent Audit Loop:
+<language>English</language>
+
+---
+
+## ⛔ STOP CHECK — Self-check before action
+
+| Am I... | Correct Action |
+|:---|:---|
+| Acting without semantic understanding first | **Stop** → Complete semantic understanding flow below first |
+| Changing requirements/logic without updating spec.md | **Stop** → Use edit flow (docs first, then code) |
+| Fixing bugs without diagnosis flow | **Stop** → Use fix flow (diagnose → Bugfix Phase → fix) |
+| Creating custom numbering (plan-01 etc.) | **Forbidden** → Use existing Task ID from roadmap.json |
+| Writing code without corresponding spec.md | **Stop** → Plan the task first |
+| Changing code for multiple tasks at once | **Stop** → Use code flow task by task |
+[[NO-COMMANDS:
+| Received `/archi.*` but didn't read protocol file | **Stop** → Read protocol full text first |
+]]
+| Modifying code without reading target file first | **Stop** → Read code first, understand context and style |
+| Executing `npx archi` without confirming working directory | **Stop** → Must be in project root (`[[__DOCS_DIR__]]/` directory) |
+| Before taking action | **Check** → Read `error_memory.json` and match `checkpoints[].before` |
+
+---
+
+## Semantic Understanding → Intent Translation → Navigation Execution
+
+**Mandatory pre-process. Forbidden to act directly after receiving user message; must complete steps below in order.**
+
+**Pre-flight (each new conversation)**: Read `[[__DOCS_DIR__]]/global/map.json`, establish cognition of project directory structure, module topology, and relationships.
+
+### Step 1: Semantic Understanding
+
+Analyze user message, extract:
+- **Action Target**: What function/module/file to operate on?
+- **Action Nature**: Create / Change / Fix / Delete / Query / Trivial edit?
+- **Scope**: Single task / Multiple tasks / Global?
+
+> One sentence may contain multiple intents → Identify all, execute in dependency order.
+
+### Step 2: Intent Translation — Layered determination of command to execute
+
+Answer one question per layer. Determine in order, stop when command is reached.
+
+**① Is this a business change?**
+No (conversation / Q&A / code reading / typo / format / comment) → Respond directly, flow ends.
+
+**② Is this initialization/tool operation?**
+Following intents bypass main flow, match directly:
+- Initialize new project → `start`
+- Adopt existing code → `inherit`
+- Manage external references → `ref`
+- Restore backup data → `recover`
+- Sync architecture map → `map`
+- Generate automation scripts → `script`
+- "What to do next" → `help`
+- None of above → Enter ③
+
+**③ Is the feature in roadmap and status allows?**
+Read `roadmap.json` to match task and check status:
+
+| Status | Action |
+|:---|:---|
+| `active` | Continue → ④ |
+| `pending` | Must `/archi.plan <ID>` first (only active tasks can be coded) |
+| `blocked` | Prompt that prerequisite dependencies are incomplete; complete dependencies first |
+| `done` | Use `/archi.edit <ID>` to modify; or revert to pending before redoing |
+| Not in roadmap | Use `scope` (must create task before planning) |
+
+**④ Is the task planned?**
+Check if `tasks/<ID>_*/spec.md` exists. Not exists → `plan` (must plan before implementing).
+
+**⑤ What does user want? — Core determination, must understand essence of request:**
+
+| User Request | Essence | Command | Prerequisite |
+|:---|:---|:---|:---|
+| "implement/build/write/develop" | Execute planned content | `code` | status=active |
+| "change/adjust/modify requirements" | Requirement/behavior/logic change | `edit` | — |
+| "fix/error/not working/bug" | Bug fix | `fix` | — |
+| "review/check quality" | Quality audit | `audit` | — |
+| "delete/remove/offline" | Feature decommission | `remove` | — |
+
+> **Ambiguity Resolution Guide** (reference when facing vague descriptions):
+> - "Change X": Changing **behavior/requirement** (spec changes) → edit; Changing **implementation detail** within spec scope → code
+> - "Optimize X": Changing **quality goals** in spec → edit; Only changing **implementation** without changing spec → code
+> - "Has issues": **Code behavior wrong** → fix; **Requirements need changing** → edit
+> - "Refactor X": **Single module** internal → edit; **Multiple modules/global** → revise
+
+**⑥ Impact scope check:**
+Changes affect **multiple tasks** or global settings (tech stack/architecture/vision) → Upgrade to `revise`.
+
+**⑦ Uncertain?**
+→ Ask user. Forbidden to guess then act.
+
+### Step 3: Navigation Execution
+
+1. **Determine command** → Read `[[__PROMPTS_PATH__]]/<protocol_file>` full text
+2. **Load context** → According to `when` conditions in file index below, only read files matching current intent; forbidden to load all
+3. **Read project code** → Before modifying code, must read target file and related files; understand context through actual code; use `map.json`'s `directoryMapping`/`logicalTopology` to locate code's corresponding spec/plan docs
+4. **Execute per protocol steps**
+
+> Code not registered in `map.json` and has no Task → **STOP & ASK**, prompt user to adopt via `/archi.inherit` or `/archi.scope`.
+
+[[NO-COMMANDS:
+## `/archi.*` Explicit Command Routing
+
+When user input starts with `/archi.`, **skip semantic understanding, load corresponding protocol directly**:
 
-  **File Metadata Protection**: When modifying files under `[[__DOCS_DIR__]]`, preserve YAML Frontmatter + `## 🤖 AI Maintenance Guide` section. Forbidden to modify or delete.
+| Command | Protocol File |
+|:---|:---|
+| `/archi.start` | `[[__PROMPTS_PATH__]]/archi.start.md` |
+| `/archi.inherit` | `[[__PROMPTS_PATH__]]/archi.inherit.md` |
+| `/archi.scope` | `[[__PROMPTS_PATH__]]/archi.scope.md` |
+| `/archi.script` | `[[__PROMPTS_PATH__]]/archi.script.md` |
+| `/archi.plan` | `[[__PROMPTS_PATH__]]/archi.plan.md` |
+| `/archi.edit` | `[[__PROMPTS_PATH__]]/archi.edit.md` |
+| `/archi.revise` | `[[__PROMPTS_PATH__]]/archi.revise.md` |
+| `/archi.code` | `[[__PROMPTS_PATH__]]/archi.code.md` |
+| `/archi.audit` | `[[__PROMPTS_PATH__]]/archi.audit.md` |
+| `/archi.fix` | `[[__PROMPTS_PATH__]]/archi.fix.md` |
+| `/archi.map` | `[[__PROMPTS_PATH__]]/archi.map.md` |
+| `/archi.remove` | `[[__PROMPTS_PATH__]]/archi.remove.md` |
+| `/archi.ref` | `[[__PROMPTS_PATH__]]/archi.ref.md` |
+| `/archi.recover` | `[[__PROMPTS_PATH__]]/archi.recover.md` |
+| `/archi.help` | `[[__PROMPTS_PATH__]]/archi.help.md` |
+
+**Protocol Load Gate** (must follow after loading protocol):
+1. **Read**: Read protocol full text → Stop if file not found
+2. **Execute**: Start from `<step_1>`, forbidden to execute subsequent steps before step_1 complete
+
+**After protocol completion**:
+
+Each protocol's Signoff contains **Next Steps** suggestion, guide user per that suggestion. Follow principles:
+
+1. **Must await user confirmation** before entering next protocol — Forbidden to auto-chain execution
+2. **Only exception**: After scope completion, may proactively ask "Want to plan the first task?"
+3. code protocol **has built-in silent audit** (step_5), no extra audit needed after completion
+4. Uncertain about next step → Suggest user run `/archi.help`
+
+> ⛔ **Special emphasis**: spec output from plan is the most important checkpoint. Forbidden to enter code without user confirming spec.
+]]
+
+---
+
+## File Index
+
+### Data Governance General Rules
+
+- **JSON Only**: `.json` under `[[__DOCS_DIR__]]/global/` is the single source of truth. `.md` views generated by `npx archi render`, forbidden to edit directly; no trailing commas, no comments.
+- **Schema Tier**: Tier 1 strict (`roadmap.json`, `plan.json`) — CLI Zod validation, forbidden to change structure arbitrarily. Tier 2 flexible (others) — only validate top-level keys, can extend fields.
+- **Read/Write Discipline**: Read `.json` not `.md` when consulting data; Prefer `npx archi task <ID> --status <s>` for Roadmap updates; Run `npx archi render` to regenerate views after updating any JSON.
+
+---
+
+### Reference Documents (read on demand)
+
+| File | Path | Contains | When to Read |
+|:---|:---|:---|:---|
+| `tech_stack.md` | `[[__DOCS_DIR__]]/global/tech_stack.md` | Tech stack selection, coding standards, naming conventions, directory structure, file placement strategy, test suite, deployment process, Anti-Patterns | When writing/modifying code; When discussing architecture/tech solutions; When creating new files to determine placement; During plan phase to assess tech feasibility; When running test/build commands |
+| `cli_reference.md` | `[[__DOCS_DIR__]]/global/references/cli_reference.md` | `npx archi` command syntax (task/plan/render), Working Directory Gate | When executing Terminal Gate in protocol Signoff; When user requests to run `npx archi` commands |
+
+---
 
-  **Post-Code Checks** (skip when: pure Q&A / no code changes / typo·comment·format only):
+### Global Data Assets (`[[__DOCS_DIR__]]/global/`)
+
+| File | Contains | When to Read | When to Write | Governance Rules |
+|:---|:---|:---|:---|:---|
+| `roadmap.json` | `[[__DOCS_DIR__]]/global/roadmap.json` | Task progress DAG: ID/title/status(pending/active/done/blocked)/deps/phase/description | During semantic understanding to match intent to task; Determine task status and deps; Check ID validity | Created by start; Appended by scope; Update status after code/fix completes | Structure: `phases[]→tasks[]`, each task must have id/title/status/deps; deps reference must exist; slug in Snake_Case |
+| `vision.md` | `[[__DOCS_DIR__]]/global/vision.md` | Project vision, North Star metrics, design philosophy, target users, boundary constraints, product principles | During plan/audit to align direction; Assess if new requirements deviate from vision | Created by start/inherit; Updated during revise | Forbidden to modify during code phase |
+| `map.json` | `[[__DOCS_DIR__]]/global/map.json` | Directory↔Module mapping (directoryMapping), logical topology (logicalTopology), user journeys (criticalUserJourneys), impact relations (featureRelations) | **Mandatory read each new conversation** (see Pre-flight); Locate code corresponding docs; Determine directory for new files; Check module dependencies; Check impact relations when modifying files | Step 3 of plan; inherit; /archi.map; **Must update immediately when creating new files/modules** | directoryMapping must reflect actual file tree; logicalTopology must register each Task Module responsibility; featureRelations organizes impact relations as networks, check other members of the same network when modifying any file; Code referencing topology violations must error and stop |
+| `dictionary.json` | `[[__DOCS_DIR__]]/global/dictionary.json` | Unified terminology: entity naming (codeName)/forbidden synonyms (forbiddenSynonyms)/verb standards/tool registry/component registry | When naming variables/classes/functions; Avoid multiple terms for same concept | Step 3 of plan registers new terms; step_5 auto-appends after code/fix | codeName is highest naming authority; Forbidden to use words in forbiddenSynonyms; Only register project business domain, forbidden to register framework concepts; UI projects only: search for reuse before creating components |
+| `error_codes.json` | `[[__DOCS_DIR__]]/global/error_codes.json` | Error code contracts: ERR_MODULE_REASON format, with message and recovery | When writing error handling code; Registering new business error codes | Step 3 of plan; step_5 auto-appends after code/fix | Format: `ERR_[MODULE]_[REASON]`; Must register before writing error handling; Only register project business errors, forbidden to register framework infrastructure errors |
+| `error_memory.json` | `[[__DOCS_DIR__]]/global/error_memory.json` | Error memory bank: keyword signature → historical solutions | Retrieve when encountering errors; Record after resolution | Append after fix succeeds; Path anonymization (relative to project root) | Prevents error loops; Auto-triggered; |
+| `design_tokens.json` | `[[__DOCS_DIR__]]/global/design_tokens.json` | UI projects only: Color palette (primitivePalette)/semantic colors (semanticTokens)/typography/border radius/spacing/motion/icon style | When writing UI code/styles | Created by start; Updated when design changes | Token Only: Styles must use Tokens, forbidden to hardcode Hex/px/rem; Must define both light and dark values |
+| `data_snapshot.json` | `[[__DOCS_DIR__]]/global/data_snapshot.json` | Data projects only: Data model snapshot: models[](name/fields/types/constraints) + relationships[](1:1/1:N/M:N) | When involving data layer design or implementation | Plan designs Schema; Sync changes after code completes | Design First: plan must specify field names and types precisely, forbidden to write TBD; Must sync back to this file after code completes |
+| `api_snapshot.json` | `[[__DOCS_DIR__]]/global/api_snapshot.json` | API projects only: API endpoint snapshot: endpoints[](route/method/params/owner) | When implementing/integrating API endpoints | Step 3 of plan registers; Sync during code | Register First: Forbidden to implement unregistered endpoints; owner marks Task ID |
+| `env_registry.json` | `[[__DOCS_DIR__]]/global/env_registry.json` | Environment variable registry: required/example/description for each env var | When introducing new config items; Checking existing env vars | Append immediately after code introduces new env var | Register when introduced; required/example mandatory |
+| `command_api.json` | `[[__DOCS_DIR__]]/global/command_api.json` | CLI projects only: CLI command registry: command/args/description/owner | When implementing/modifying CLI commands | Step 3 of plan registers; Sync during code | Sync when modified; owner marks Task ID |
+| `public_api.json` | `[[__DOCS_DIR__]]/global/public_api.json` | Lib projects only: Library export registry: complete TS signature/stability marker for exports | When adding/modifying public APIs | Step 3 of plan registers; Sync during code | stable changes must go through /archi.edit; Complete TS signature; owner marks Task ID |
+
+---
+
+### Task Documents (`[[__DOCS_DIR__]]/tasks/<ID>_<Slug>/`)
+
+| File | Contains | When to Read | When to Write | Governance Rules |
+|:---|:---|:---|:---|:---|
+| `spec.md` | Feature spec: Gherkin scenarios (Feature)/config contracts (Infra)/interface definitions/constraints (Constraints) | Before writing code to understand requirements; Check Spec drift | Created by plan; Modified by edit | Forbidden to modify spec during code phase; If spec insufficient found after code, must go through edit |
+| `plan.json` | Implementation plan: Phase/Task breakdown, done markers, decisions (with rationale), notes (with verification method) | During coding per Phase; Check completion status | Created by plan; Set done to true after code step completes; Append new Phases during edit while preserving completed history | Tier 1 strict; Append Phase forbidden to overwrite completed history |
+| `ui.md` | UI projects only: Task-level UI scope declaration: reference screen ID in ui_context.md, delta components | When writing UI code | Created by plan; Modified by edit | Forbidden to redefine global layout; Must reference screen ID registered in ui_context.md |
+| `design.md` | Complex tasks only: Core mechanism technical solution: state machines/pipelines/protocol definitions, parameter tables, invariants, failure modes | When implementing complex logic | Created by plan (optional) | Forbidden to modify during code phase |
+| `audit.md` | Audit report: findings list (CRITICAL/WARNING/INFO), fix tickets | When viewing audit results; Before fix to understand known issues | Created by audit; Update status after fix completes | Forbidden to directly modify findings list; Update status in ticket after fix |
+
+---
+
+## Post-Action: Self-check after code output
+
+> Skip conditions: Pure Q&A / No code changes / Only typo·comment·formatting
+
+**A. Spec Drift** (when spec.md was loaded):
+- ✅ Change within spec scope → No action needed
+- ⚠️ Exceeds spec scope (new interface·changed signature·new behavior·new scenario) → Output `⚠️ Spec Drift`, suggest `/archi.edit <ID>`
+
+**B. Data Governance**:
+When this change introduces new modules/entities/error codes/endpoints/commands/exports/design tokens/env vars etc. global data assets → Execute incremental sync per **write conditions** and **governance rules** in file index above for corresponding files.
+
+**C. File Metadata Protection**:
+When modifying files under `[[__DOCS_DIR__]]`, preserve YAML Frontmatter + `## 🤖 AI Maintenance Guide` section, forbidden to modify or delete.
+
+**D. Map Maintenance**:
+1. **Before modifying**: Check `map.json` for `featureRelations` (impact associations) and `logicalTopology` (module dependencies), confirm if any related files need to be processed together
+2. **After modifying**:
+   - New directories/modules → Update `directoryMapping`
+   - New module dependencies → Update `logicalTopology`
+   - New impact relations → Update `featureRelations`
+   - New user journeys → Update `criticalUserJourneys`
+3. Only ask user confirmation when mapping relationship is unclear
+
+> Core principle: **When modifying one file, automatically check and sync related map.json fields**.
+
+---
 
-  **A. Spec Drift** (when spec.md was loaded):
-  - ✅ Change within spec scope → No action needed
-  - ⚠️ Exceeds spec scope (new interface · changed signature · new behavior · new scenario) → Output `⚠️ Spec Drift`; recommend `/archi.edit <ID>`
+## Error Memory — Prevent Error Loops
 
-  **B. Data Governance**:
-  [[SUBAGENT: archi-data-sync|context: Scan changes for new business entities/error codes/Schema, incrementally sync per 03_data_governance.md rules]][[NO-SKILL: When changes introduce new business entities/error codes/Schema, execute incremental sync per `03_data_governance.md` rules.]]
-</thinking_process>
+### Prediction Phase (Before Execution)
 
-<communication_style>
-  <language>English</language>
+## Error Memory — Prevent Error Loops
 
-  <safety>
-When involving Schema Change / File Deletion / Dependency Install, must list changes and request confirmation.
-</safety>
-</communication_style>
+**Prediction** (before execution): Match `checkpoints[].before` → Output `watchFor` + `lesson` reminder
+**Matching** (after error): Match `errorPatterns[].matchWhen` → Output `cause` + `solution` + `lesson`
+**Recording** (after resolution): Append `errorPatterns[]{id,watchFor,matchWhen,cause,solution,lesson}` + `checkpoints[]{before,check}`

package/dist/templates/en/rules/90_custom_rules.md

@@ -8,10 +8,12 @@ alwaysApply: true
 # Custom Project Rules
 
 > **Priority:** [High]
-> **Role:** The "House Rules" of this project. When general `02_tech_stack` cannot cover certain details, or there are specific business constraints, this file prevails.
+> **Role:** The "House Rules" of this project. When general `tech_stack` cannot cover certain details, or there are specific business constraints, this file prevails.
 
 ## 1. Team Conventions
 
+- **Git Commit**: Provide commit suggestions (in Simplified Chinese, following Conventional Commits) after each code or doc change; default not to commit automatically, confirm by user.
+
 
 ## 2. Business Constraints
 

package/dist/templates/en/skills/archi-data-sync/SKILL.md

@@ -1,7 +1,6 @@
 ---
 name: archi-data-sync
-type: reviewer
-description: Data governance sync executor. In isolated context, scans main Agent output, compares with global data file state, performs incremental sync per 03_data_governance.md rules, returns change Diff.
+description: Sync data governance files with main agent output. **Must run in isolated context/subagent.** Use when verifying data consistency between code and global data files.
 ---
 
 # Data Governance Sync Executor
@@ -18,14 +17,14 @@ Main Agent Signoff (confirm Diff)
 
 > **Skill responsibility boundary**:
 > - Responsible: Scan main Agent output for new business entities/error codes/Schema, compare with global data files, perform incremental sync
-> - Not responsible: Modify `03_data_governance.md` itself (this Skill enforces rules, does not define them)
+> - Not responsible: Modify `00_system.md` itself (this Skill enforces rules, does not define them)
 > - Not responsible: Register framework concepts (Architext framework concepts must not be registered to global data files)
 
 ---
 
 ## Authoritative Rule Source
 
-`03_data_governance.md` is the single authoritative source for data governance. All behavior of this Skill must align with that file.
+`00_system.md` "File Index → Global Data Assets" table is the authoritative rule source for data governance. All behavior of this Skill must align with that file.
 
 ---
 
@@ -33,11 +32,14 @@ Main Agent Signoff (confirm Diff)
 
 | Global file | Sync content | Trigger |
 |:---|:---|:---|
+| `map.json` | New module registration (directoryMapping), dependency relationships (logicalTopology), impact associations (featureRelations) | When creating new code modules/directories |
 | `dictionary.json` | New business entities · actions · shared tools · public components | Output contains unregistered business terms or tools |
 | `error_codes.json` | New business error codes | Output contains unregistered error scenarios |
+| `env_registry.json` | New env vars | Output introduces new `process.env.X` |
+| [?UI] `design_tokens.json` | Style changes | Output has new color/font/spacing/motion definitions |
+| [?UI] `ui_context.md` | Screen index changes | Output has new/modified screens |
 | [?Data] `data_snapshot.json` | Schema changes | Output has data model add/modify |
 | [?API] `api_snapshot.json` | New endpoints | Output has new HTTP/RPC endpoints |
-| [?API] `env_registry.json` | New env vars | Output introduces new `process.env.X` |
 | [?CLI] `command_api.json` | New commands | Output has new CLI commands |
 | [?Lib] `public_api.json` | New public exports | Output has new public exports |
 
@@ -45,18 +47,25 @@ Main Agent Signoff (confirm Diff)
 
 ## Execution Protocol
 
-1. **Read global data files**: Load global files matching project features from table above
-2. **Scan main Agent output**: Identify new business entities, error codes, Schema, endpoints, etc.
-3. **Boundary check**: Do not register framework concepts (scripts, scaffold, roadmap, plan, etc.); sync only project business domain content
-4. **Deduplication**: Compare with existing entries, avoid duplicate registration
-5. **Incremental sync**: Append/modify only, do not delete existing entries
-6. **Output change Diff**
+1. **Read project features**: Read project root `architext.json`, extract `features` field (api/cli/lib/data/ui etc.)
+2. **Read global data files**: Load global files matching project features from table above
+3. **Scan main Agent output**: Identify changes based on file type:
+   - **Code files**: New modules → map.json, new types/interfaces → dictionary.json, new error handling → error_codes.json, new env vars → env_registry.json
+   - **UI code**: New colors/fonts/components → design_tokens.json, new routes/screens → ui_context.md
+   - **Data code**: New models/fields → data_snapshot.json
+   - **API code**: New endpoints → api_snapshot.json
+   - **CLI code**: New commands → command_api.json
+   - **Lib code**: New exports → public_api.json
+4. **Boundary check**: Do not register framework concepts (scripts, scaffold, roadmap, plan, etc.); sync only project business domain content
+5. **Deduplication**: Compare with existing entries, avoid duplicate registration
+6. **Incremental sync**: Append/modify only, do not delete existing entries
+7. **Output change Diff**
 
 ### Hard Boundaries
 
 - **No direct append** — Must check existing content boundary before write, avoid duplicates or conflicts
 - **No framework concept registration** — Sync only project business domain content
-- **No modify `03_data_governance.md`** — This Skill enforces rules, does not define them
+- **No modify `00_system.md`** — This Skill enforces rules, does not define them
 
 ### Output Format
 
@@ -81,3 +90,8 @@ When no sync needed: `### Data Sync Results — NO CHANGES`
 ---
 
 > **Intermediate output**: This Skill is a review subprogram; after producing change Diff, control returns to caller.
+
+## Output Verification
+
+□ Target global JSON files updated (dictionary.json / error_codes.json / data_snapshot.json / etc.)
+□ Diff output generated with ADDED/MODIFIED/NO CHANGE sections