dw-kit 1.3.6 → 1.6.0-rc.1

package/.dw/adapters/generic/AGENT.md

@@ -1,169 +1,171 @@
-# dw-kit Workflow — Generic Agent Instructions
-
-> **For**: Cursor, Windsurf, Copilot Chat, or any AI coding assistant
-> **Note**: This is a methodology reference. It cannot replicate Claude Code-specific features
-> (agent delegation with tool constraints, hooks, MCP integration). For full capabilities, use the Claude CLI adapter.
-
----
-
-## Setup
-
-1. Copy this file to your project root as `AGENT.md` (or paste into your AI assistant's context)
-2. Copy `.dw/config/dw.config.yml` to your project and fill in your settings
-3. Create `.dw/tasks/` directory for task documentation
-
----
-
-## Configuration
-
-Read `.dw/config/dw.config.yml` for:
-- `workflow.default_depth` — quick | standard | thorough
-- `team.roles` — available team roles
-- `quality.test_command`, `quality.lint_command` — quality gates
-- `paths.tasks` — where task docs are stored
-
----
-
-## Depth Routing
-
-| Scope | Depth | Workflow |
-|-------|-------|---------|
-| ≤2 files, hotfix | quick | Understand → Execute → Close |
-| 3-5 files | standard | All 6 phases |
-| 6+ files, API/DB changes | thorough | All phases + arch review |
-
----
-
-## Workflow Phases
-
-### Phase 1: Initialize
-
-Create task documentation at `{paths.tasks}/[task-name]/`:
-```
-[name]-context.md    # Research findings
-[name]-plan.md       # Implementation plan
-[name]-progress.md   # Progress tracking
-```
-
-**Before proceeding**: Task name defined, scope assessed, depth chosen.
-
-### Phase 2: Understand (Research)
-
-Explore the codebase:
-- Find all files related to the task
-- Map dependencies (upstream and downstream)
-- Identify current patterns and conventions
-- Check test coverage
-- Note git history for recent changes in the area
-- Document what is unclear
-
-Fill in `[name]-context.md` with findings.
-
-**Before proceeding**: All files identified, no critical unknowns.
-
-### Phase 3: Plan
-
-Design solution before writing any code:
-- Consider ≥2 approaches with trade-offs
-- Apply devil's advocate: strongest reason NOT to choose your preferred approach
-- Break into subtasks (each ≤3 files, ≤4 hours, independent commit)
-- Order: schema → service → API → tests → docs
-
-Fill in `[name]-plan.md`.
-
-**STOP**: Wait for human approval before executing.
-If team has TL: TL must review architecture decisions.
-
-### Phase 4: Execute
-
-For each subtask:
-1. Read acceptance criteria
-2. Write tests first (failing) — RED
-3. Implement to make tests pass — GREEN
-4. Refactor if needed — REFACTOR
-5. Update progress file
-6. Commit: `type(scope): description`
-
-**Rules**:
-- Only work within subtask scope
-- If ambiguous → stop and ask (don't guess for large changes)
-- If scope changes → update plan, ask human
-
-### Phase 5: Verify
-
-**Self-review** (always):
-- Logic correct? Edge cases handled?
-- No debug code (console.log, debugger, etc.)
-- No sensitive data (passwords, tokens, keys)
-
-**Automated** (if configured):
-```bash
-{quality.test_command}
-{quality.lint_command}
-```
-
-**Peer/TL review** (standard+):
-- Architecture decisions reviewed
-- Code review with checklist (see `.dw/core/QUALITY.md`)
-
-**QA** (thorough + qc role):
-- QA reviews against test plan
-- Explicit sign-off required
-
-### Phase 6: Close
-
-Commit format:
-```
-<type>(<scope>): <description ≤72 chars>
-
-Co-Authored-By: [AI assistant name]
-```
-
-Update progress file: status → Done.
-If handing off: write handoff notes (done/in-progress/blocked, next steps).
-
----
-
-## Standalone: Debug
-
-When encountering a bug:
-1. **Investigate**: Reproduce exactly. Gather evidence (error, stack trace, file:line).
-2. **Diagnose**: Form hypothesis. Verify. Find root cause (not symptom).
-3. **Fix**: Fix root cause. Test the fix. Check for regressions. Commit.
-
----
-
-## Quality Principles
-
-Full quality strategy: `.dw/core/QUALITY.md`
-
-Key principles:
-- Write tests before implementation (TDD)
-- Each subtask = one commit
-- Review checklist: correctness, security, performance, maintainability, tests
-- CRITICAL issues must be fixed before merge
-
----
-
-## Limitations of This Adapter
-
-This generic adapter provides the **methodology** but cannot provide:
-
-| Feature | Claude CLI Adapter | Generic Adapter |
-|---------|-------------------|-----------------|
-| Agent delegation (researcher/planner/reviewer) | ✅ | ❌ |
-| Tool constraints (read-only research agent) | ✅ | ❌ |
-| Pre-commit hooks (safety-guard, quality gates) | ✅ | ❌ |
-| MCP server integration | ✅ | ❌ |
-| Automatic progress tracking | ✅ | Manual |
-
-For full capabilities, use Claude Code with the Claude CLI adapter.
-
----
-
-## Full Methodology Reference
-
-- Workflow phases: `.dw/core/WORKFLOW.md`
-- Thinking framework: `.dw/core/THINKING.md`
-- Quality strategy: `.dw/core/QUALITY.md`
-- Role definitions: `.dw/core/ROLES.md`
+# dw-kit Workflow — Generic Agent Instructions
+
+> **For**: Cursor, Windsurf, Copilot Chat, or any AI coding assistant
+> **Note**: This is a methodology reference. It cannot replicate Claude Code-specific features
+> (agent delegation with tool constraints, hooks, MCP integration). For full capabilities, use the Claude CLI adapter.
+
+---
+
+## Setup
+
+1. Copy this file to your project root as `AGENT.md` (or paste into your AI assistant's context)
+2. Copy `.dw/config/dw.config.yml` to your project and fill in your settings
+3. Create `.dw/tasks/` directory for task documentation
+
+---
+
+## Configuration
+
+Read `.dw/config/dw.config.yml` for:
+- `workflow.default_depth` — quick | standard | thorough
+- `team.roles` — available team roles
+- `quality.test_command`, `quality.lint_command` — quality gates
+- `paths.tasks` — where task docs are stored
+
+---
+
+## Depth Routing
+
+| Scope | Depth | Workflow |
+|-------|-------|---------|
+| ≤2 files, hotfix | quick | Understand → Execute → Close |
+| 3-5 files | standard | All 6 phases |
+| 6+ files, API/DB changes | thorough | All phases + arch review |
+
+---
+
+## Workflow Phases
+
+### Phase 1: Initialize
+
+Create task documentation at `{paths.tasks}/[task-name]/`:
+```
+[name]-context.md    # Research findings
+[name]-plan.md       # Implementation plan
+[name]-progress.md   # Progress tracking
+```
+
+**Before proceeding**: Task name defined, scope assessed, depth chosen.
+
+### Phase 2: Understand (Research)
+
+Explore the codebase:
+- Find all files related to the task
+- Map dependencies (upstream and downstream)
+- Identify current patterns and conventions
+- Check test coverage
+- Note git history for recent changes in the area
+- Document what is unclear
+
+Fill in `[name]-context.md` with findings.
+
+**Before proceeding**: All files identified, no critical unknowns.
+
+### Phase 3: Plan
+
+Design solution before writing any code:
+- Consider ≥2 approaches with trade-offs
+- Apply devil's advocate: strongest reason NOT to choose your preferred approach
+- Break into subtasks (each ≤3 files, ≤4 hours, independent commit)
+- Order: schema → service → API → tests → docs
+
+Fill in `[name]-plan.md`.
+
+**STOP**: Wait for human approval before executing.
+If team has TL: TL must review architecture decisions.
+
+### Phase 4: Execute
+
+For each subtask:
+1. Read acceptance criteria
+2. Write tests first (failing) — RED
+3. Implement to make tests pass — GREEN
+4. Refactor if needed — REFACTOR
+5. Update progress file
+6. Commit: `type(scope): description`
+
+**Rules**:
+- Only work within subtask scope
+- If ambiguous → stop and ask (don't guess for large changes)
+- If scope changes → update plan, ask human
+
+### Phase 5: Verify
+
+**Self-review** (always):
+- Logic correct? Edge cases handled?
+- No debug code (console.log, debugger, etc.)
+- No sensitive data (passwords, tokens, keys)
+
+**Automated** (if configured):
+```bash
+{quality.test_command}
+{quality.lint_command}
+```
+
+**Peer/TL review** (standard+):
+- Architecture decisions reviewed
+- Code review with checklist (see `.dw/core/QUALITY.md`)
+
+**QA** (thorough + qc role):
+- QA reviews against test plan
+- Explicit sign-off required
+
+### Phase 6: Close
+
+Commit format:
+```
+<type>(<scope>): <imperative English subject ≤72 chars>
+
+[Optional body — explain WHY, wrap at 72]
+```
+
+English imperative mood. Do not append `Co-Authored-By` AI signatures. Full rules: `.claude/rules/commit-standards.md`.
+
+Update progress file: status → Done.
+If handing off: write handoff notes (done/in-progress/blocked, next steps).
+
+---
+
+## Standalone: Debug
+
+When encountering a bug:
+1. **Investigate**: Reproduce exactly. Gather evidence (error, stack trace, file:line).
+2. **Diagnose**: Form hypothesis. Verify. Find root cause (not symptom).
+3. **Fix**: Fix root cause. Test the fix. Check for regressions. Commit.
+
+---
+
+## Quality Principles
+
+Full quality strategy: `.dw/core/QUALITY.md`
+
+Key principles:
+- Write tests before implementation (TDD)
+- Each subtask = one commit
+- Review checklist: correctness, security, performance, maintainability, tests
+- CRITICAL issues must be fixed before merge
+
+---
+
+## Limitations of This Adapter
+
+This generic adapter provides the **methodology** but cannot provide:
+
+| Feature | Claude CLI Adapter | Generic Adapter |
+|---------|-------------------|-----------------|
+| Agent delegation (researcher/planner/reviewer) | ✅ | ❌ |
+| Tool constraints (read-only research agent) | ✅ | ❌ |
+| Pre-commit hooks (safety-guard, quality gates) | ✅ | ❌ |
+| MCP server integration | ✅ | ❌ |
+| Automatic progress tracking | ✅ | Manual |
+
+For full capabilities, use Claude Code with the Claude CLI adapter.
+
+---
+
+## Full Methodology Reference
+
+- Workflow phases: `.dw/core/WORKFLOW.md`
+- Thinking framework: `.dw/core/THINKING.md`
+- Quality strategy: `.dw/core/QUALITY.md`
+- Role definitions: `.dw/core/ROLES.md`

package/.dw/config/config.schema.json

@@ -1,121 +1,149 @@
-{
-  "$schema": "http://json-schema.org/draft-07/schema#",
-  "title": "dw-kit Configuration Schema",
-  "description": "Schema for config/dw.config.yml",
-  "type": "object",
-  "additionalProperties": false,
-  "properties": {
-    "project": {
-      "type": "object",
-      "additionalProperties": false,
-      "required": ["name"],
-      "properties": {
-        "name":     { "type": "string", "minLength": 1 },
-        "language": { "type": "string", "enum": ["vi", "en"], "default": "vi" }
-      }
-    },
-    "workflow": {
-      "type": "object",
-      "additionalProperties": false,
-      "properties": {
-        "default_depth": {
-          "type": "string",
-          "enum": ["quick", "standard", "thorough"],
-          "default": "standard"
-        }
-      }
-    },
-    "team": {
-      "type": "object",
-      "additionalProperties": false,
-      "properties": {
-        "roles": {
-          "type": "array",
-          "items": {
-            "type": "string",
-            "enum": ["dev", "techlead", "ba", "qc", "pm"]
-          },
-          "minItems": 1,
-          "contains": { "const": "dev" },
-          "description": "dev is always required"
-        }
-      }
-    },
-    "quality": {
-      "type": "object",
-      "additionalProperties": false,
-      "properties": {
-        "test_command":  { "type": "string", "default": "" },
-        "lint_command":  { "type": "string", "default": "" },
-        "block_on_fail": { "type": "boolean", "default": false }
-      }
-    },
-    "tracking": {
-      "type": "object",
-      "additionalProperties": false,
-      "properties": {
-        "estimation":      { "type": "boolean", "default": false },
-        "log_work":        { "type": "boolean", "default": false },
-        "estimation_unit": {
-          "type": "string",
-          "enum": ["hours", "story-points", "t-shirt"],
-          "default": "hours"
-        }
-      }
-    },
-    "paths": {
-      "type": "object",
-      "additionalProperties": false,
-      "properties": {
-        "tasks": { "type": "string", "default": ".dw/tasks" },
-        "docs":  { "type": "string", "default": ".dw/docs" }
-      }
-    },
-    "claude": {
-      "type": "object",
-      "additionalProperties": false,
-      "properties": {
-        "models": {
-          "type": "object",
-          "additionalProperties": false,
-          "properties": {
-            "plan":    { "type": "string", "default": "" },
-            "execute": { "type": "string", "default": "" },
-            "review":  { "type": "string", "default": "" }
-          }
-        },
-        "structured_output":  { "type": "boolean", "default": true },
-        "worktree_execution": { "type": "boolean", "default": false },
-        "mcp": {
-          "type": "array",
-          "items": {
-            "type": "object",
-            "required": ["name", "command"],
-            "additionalProperties": false,
-            "properties": {
-              "name":    { "type": "string" },
-              "command": { "type": "string" },
-              "args":    { "type": "array", "items": { "type": "string" } },
-              "env": {
-                "type": "object",
-                "additionalProperties": { "type": "string" }
-              }
-            }
-          }
-        }
-      }
-    },
-    "_toolkit": {
-      "type": "object",
-      "additionalProperties": true,
-      "properties": {
-        "core_version":     { "type": "string" },
-        "platform_version": { "type": "string" },
-        "capability_version": { "type": "string" },
-        "installed":        { "type": "string" },
-        "last_upgrade":     { "type": "string" },
-        "migrated_from":    { "type": "string" }
-      }
-    }
-  }
-}
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "title": "dw-kit Configuration Schema",
+  "description": "Schema for config/dw.config.yml",
+  "type": "object",
+  "additionalProperties": false,
+  "properties": {
+    "project": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["name"],
+      "properties": {
+        "name":     { "type": "string", "minLength": 1 },
+        "language": { "type": "string", "enum": ["vi", "en"], "default": "vi" }
+      }
+    },
+    "workflow": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "default_depth": {
+          "type": "string",
+          "enum": ["quick", "standard", "thorough"],
+          "default": "standard"
+        }
+      }
+    },
+    "team": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "roles": {
+          "type": "array",
+          "items": {
+            "type": "string",
+            "enum": ["dev", "techlead", "ba", "qc", "pm"]
+          },
+          "minItems": 1,
+          "contains": { "const": "dev" },
+          "description": "dev is always required"
+        }
+      }
+    },
+    "quality": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "test_command":  { "type": "string", "default": "" },
+        "lint_command":  { "type": "string", "default": "" },
+        "block_on_fail": { "type": "boolean", "default": false }
+      }
+    },
+    "tracking": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "estimation":      { "type": "boolean", "default": false },
+        "log_work":        { "type": "boolean", "default": false },
+        "estimation_unit": {
+          "type": "string",
+          "enum": ["hours", "story-points", "t-shirt"],
+          "default": "hours"
+        }
+      }
+    },
+    "paths": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "tasks": { "type": "string", "default": ".dw/tasks" },
+        "docs":  { "type": "string", "default": ".dw/docs" }
+      }
+    },
+    "claude": {
+      "type": "object",
+      "additionalProperties": false,
+      "properties": {
+        "models": {
+          "type": "object",
+          "additionalProperties": false,
+          "properties": {
+            "plan":    { "type": "string", "default": "" },
+            "execute": { "type": "string", "default": "" },
+            "review":  { "type": "string", "default": "" }
+          }
+        },
+        "structured_output":  { "type": "boolean", "default": true },
+        "worktree_execution": { "type": "boolean", "default": false },
+        "review": {
+          "type": "object",
+          "additionalProperties": false,
+          "description": "Review render pipeline (ADR-0007).",
+          "properties": {
+            "renderer": {
+              "type": "object",
+              "additionalProperties": false,
+              "properties": {
+                "strategy": {
+                  "type": "string",
+                  "enum": ["auto", "plugin", "markdown-only"],
+                  "default": "auto"
+                },
+                "theme":  { "type": "string", "default": "github-dark" },
+                "font":   { "type": "string", "default": "JetBrains Mono" },
+                "formats": {
+                  "type": "array",
+                  "items": { "type": "string", "enum": ["svg", "png"] },
+                  "minItems": 1,
+                  "uniqueItems": true,
+                  "default": ["svg", "png"]
+                },
+                "output_dir": { "type": "string", "default": ".dw/reviews" }
+              }
+            }
+          }
+        },
+        "mcp": {
+          "type": "array",
+          "items": {
+            "type": "object",
+            "required": ["name", "command"],
+            "additionalProperties": false,
+            "properties": {
+              "name":    { "type": "string" },
+              "command": { "type": "string" },
+              "args":    { "type": "array", "items": { "type": "string" } },
+              "env": {
+                "type": "object",
+                "additionalProperties": { "type": "string" }
+              }
+            }
+          }
+        }
+      }
+    },
+    "_toolkit": {
+      "type": "object",
+      "additionalProperties": true,
+      "properties": {
+        "core_version":     { "type": "string" },
+        "platform_version": { "type": "string" },
+        "capability_version": { "type": "string" },
+        "installed":        { "type": "string" },
+        "last_upgrade":     { "type": "string" },
+        "migrated_from":    { "type": "string" }
+      }
+    }
+  }
+}

package/.dw/config/dw.config.yml

@@ -59,6 +59,20 @@ claude:
   # Isolate execution trong git worktree (cho risky refactors)
   worktree_execution: false
 
+  # --- Review render pipeline (ADR-0007) ------------------------------------
+  # /dw:review --visual produces a manifest.json then invokes `dw review render`.
+  # Output goes to .dw/reviews/{scope-slug}/ — see ADR-0007.
+  review:
+    renderer:
+      # auto: try dw-kit-render package, fall back to markdown summary if missing
+      # plugin: require dw-kit-render to be installed (fail if missing)
+      # markdown-only: never invoke a renderer, emit markdown summary only
+      strategy: "auto"
+      theme: "github-dark"          # any shiki theme name
+      font: "JetBrains Mono"        # font family for code in SVG
+      formats: ["svg", "png"]       # outputs produced per finding
+      output_dir: ".dw/reviews"
+
   # MCP servers — Claude Code sẽ load các servers này
   mcp: []
   # Ví dụ: