@fitlab-ai/agent-infra 0.6.5 → 0.7.0

package/templates/.agents/skills/review-plan/reference/report-template.en.md

@@ -0,0 +1,90 @@
+# Review Report Template
+
+Use this template when writing `review-plan.md` or `review-plan-r{N}.md`.
+
+## Output Template
+
+```markdown
+# Technical Plan Review Report
+
+- **Review Round**: Round {review-round}
+- **Artifact File**: `{review-artifact}`
+- **Review Input**:
+  - `{plan-artifact}`
+
+## State Check
+
+> Paste the raw state-check command output; each command starts with `$ `.
+
+## Review Summary
+
+- **Reviewer**: {reviewer-name}
+- **Review Time**: {timestamp}
+- **Scope**: {file-count and major modules}
+- **Overall Verdict**: {Approved / Changes Requested / Rejected}
+- **Findings (AI-actionable)**: 0 blockers, 0 majors, 0 minors / **env-blocked**: 0
+
+## Findings
+
+### Blockers (must fix)
+
+#### 1. {Issue title}
+**File**: `{file-path}:{line-number}`
+**Description**: {details}
+**Suggested Fix**: {fix suggestion}
+
+### Major Issues (should fix)
+
+#### 1. {Issue title}
+**File**: `{file-path}:{line-number}`
+**Description**: {details}
+**Suggested Fix**: {fix suggestion}
+
+### Minor Issues (optional improvements)
+
+#### 1. {Improvement point}
+**File**: `{file-path}:{line-number}`
+**Suggestion**: {improvement suggestion}
+
+## Environment-Blocked Findings
+
+> Items the AI agent cannot close in the current execution environment; they do not participate in the next plan round. Maintainers carry them in the PR description as a "manual verification required" checklist.
+
+#### 1. {environment-blocked finding title}
+**File**: `{file-path}:{line-number}` (if applicable)
+**Description**: {details}
+**Required Environment**: {e.g. Docker sandbox / macOS host / privileged root / third-party account}
+**Manual Verification Steps**: {steps for the human verifier}
+
+> If this round has no env-blocked findings, keep the section heading and write "None".
+
+
+## Evidence
+
+> Pair each "I verified X" claim with the corresponding raw tool output; the gate only checks that this section exists and at least one `$ ` line is present.
+
+- Claim: {verified claim}
+```text
+$ {command}
+{raw output}
+```
+
+## Highlights
+
+- {what went well}
+
+## Alignment with Plan
+
+- [ ] Implementation matches the technical plan
+- [ ] No unintended scope expansion
+
+## Conclusion and Recommendation
+
+### Approval Decision
+- [ ] Approved
+- [ ] Changes Requested
+- [ ] Rejected
+
+### Next Steps
+{recommended next step}
+```

package/templates/.agents/skills/{review-task → review-plan}/reference/report-template.zh-CN.md

@@ -1,6 +1,6 @@
 # 审查报告模板
 
-编写 `review.md` 或 `review-r{N}.md` 时使用本模板。
+编写 `review-plan.md` 或 `review-plan-r{N}.md` 时使用本模板。
 
 ## 输出模板
 
@@ -10,8 +10,8 @@
 - **审查轮次**：第 {review-round} 轮
 - **产物文件**：`{review-artifact}`
 - **审查输入**：
-  - `{implementation-artifact}`
-  - `{refinement-artifact}`（如存在）
+  - `{code-artifact}`
+  - `{code-artifact}`（如存在）
 
 ## 状态核对
 

package/templates/.agents/skills/review-plan/reference/review-criteria.en.md

@@ -0,0 +1,47 @@
+# Review Criteria
+
+Read this file before reviewing technical plan artifacts or classifying finding severity.
+
+## Technical Plan Review
+
+Follow the `design-review` step in `.agents/workflows/feature-development.yaml`.
+
+**Required review areas**:
+- [ ] The plan covers the approved requirement analysis
+- [ ] Implementation steps are concrete, ordered, and verifiable
+- [ ] Architecture boundaries, data flow, and interface changes are clear
+- [ ] Test strategy covers critical paths, regression risks, and edge cases
+- [ ] Risks, migration, rollback, or compatibility handling are sufficient
+- [ ] The plan avoids over-design and unrelated scope expansion
+
+**Common anti-examples**:
+- Saying "modify related code" without executable steps and verification points
+- Ignoring risks or constraints listed in the analysis
+- Introducing unnecessary abstractions, configuration, or frameworks for a single-use requirement
+
+## Common Review Principles
+
+1. **Strict but fair**: identify issues and acknowledge solid work
+2. **Specific**: cite exact file paths and line numbers
+3. **Actionable**: suggest a concrete fix
+4. **Severity-based**: clearly distinguish blockers, major issues, and minor issues
+
+## Environment-Blocked Classification
+
+Some findings cannot be closed by an AI agent in the current execution environment, for example:
+
+- Missing Docker / sandbox access for end-to-end validation
+- Missing a specific OS for macOS-only behavior
+- Missing third-party accounts / OAuth
+- Missing privileged operations such as root, sudo, or special network access
+
+**Decision tree**: "Can the AI agent close this item independently without changing the environment?"
+- Yes -> blocker / major / minor, based on risk
+- No -> **env-blocked** (a meta-category, not part of severity ordering)
+
+Where env-blocked items go:
+- Record them in an independent review report section named "Environment-Blocked Findings"
+- Include them at the end of the numeric summary, for example `(+ 1 env-blocked)`
+- Do **not** include them in the code-task fix loop; maintainers carry them in the PR description under manual verification
+
+Also inspect the latest technical plan artifact, latest requirement-analysis review artifact, and `task.md` Activity Log so the report reflects the full design context.

package/templates/.agents/skills/review-plan/reference/review-criteria.zh-CN.md

@@ -0,0 +1,47 @@
+# 审查标准
+
+在审查技术方案或划分问题严重程度之前先读取本文件。
+
+## 执行技术方案审查
+
+遵循 `.agents/workflows/feature-development.yaml` 中的 `design-review` 步骤。
+
+**必查范围**：
+- [ ] 方案是否覆盖已批准的需求分析
+- [ ] 实现步骤是否具体、顺序合理且可验证
+- [ ] 架构边界、数据流和接口变化是否清晰
+- [ ] 测试策略是否覆盖关键路径、回归风险和边界情况
+- [ ] 风险、迁移、回滚或兼容性处理是否充分
+- [ ] 方案是否避免过度设计和无关扩张
+
+**常见反例**：
+- 方案只写“修改相关代码”，没有可执行步骤和验证点
+- 设计没有回应分析中列出的风险或约束
+- 为单次需求引入不必要的新抽象、配置或框架
+
+## 通用审查原则
+
+1. **严格但公正**：既要指出问题，也要承认做得好的部分
+2. **具体**：引用准确的文件路径和行号
+3. **可执行**：给出明确可落地的修复建议
+4. **按严重程度分类**：明确区分 blocker、major 和 minor
+
+## 环境性遗留分类
+
+某些发现项是 AI agent 在本执行环境**无法闭环**的，例如：
+
+- 缺 Docker / 沙箱而无法跑端到端验证
+- 缺特定 OS（macOS-only 行为）
+- 缺第三方账号 / OAuth
+- 缺特权操作（root、sudo、特殊网络）
+
+**分类决策树**：「AI agent 能否在不改环境的前提下独立闭环这一项？」
+- 是 -> blocker / major / minor 之一（按风险定档）
+- 否 -> **env-blocked**（元类目，不参与严重程度排序）
+
+env-blocked 项的去向：
+- 写入 review 报告独立段落「环境性遗留」
+- 在数字摘要末尾附带显示（如 `(+ 1 env-blocked)`）
+- **不**进入 code-task 修复循环；维护者在 PR description 中以「待人工验证」清单形式承接
+
+同时检查最新技术方案产物、最新需求分析审查产物和 `task.md` Activity Log，确保报告反映完整的设计上下文。

package/templates/.agents/skills/test/SKILL.en.md

@@ -35,7 +35,7 @@ This project uses three test layers as an optional optimization; if the test sui
 ```
 
 Use for:
-- implement-task / refine-task inner loops
+- code-task inner loops
 - save-and-run / frequent feedback
 - project structure, configuration, and template contract checks
 
@@ -50,7 +50,7 @@ Use for:
 
 Use for:
 - pre-commit hook (automatic)
-- final verification before writing implementation.md / refinement.md
+- final verification before writing code.md / code-r{N}.md
 - local gate before pushing a PR
 
 ### full (complete test suite)

package/templates/.agents/skills/test/SKILL.zh-CN.md

@@ -1,74 +1,56 @@
 ---
 name: test
-description: "执行项目完整测试流程"
+description: >
+  执行项目完整测试流程（编译检查 + 单元测试）。
+  当用户要求运行测试或验证代码质量时触发。
 ---
 
 # 执行测试
 
 执行项目的完整测试流程，包括编译检查和单元测试。
 
-<!-- TODO: 将以下命令替换为你的项目实际命令 -->
-
 ## 1. 编译 / 类型检查
 
-```bash
-# TODO: 替换为你的项目编译命令
-# npx tsc --noEmit       (TypeScript)
-# mvn compile             (Maven)
-# go build ./...          (Go)
-# make build              (通用)
-```
-
-确认无编译错误。
+本项目由 Node.js CLI 和模板文件组成，无需编译。跳过此步骤。
 
 ## 2. 运行单元测试（按层级选择）
 
-本项目把测试分为三层（可选优化）；如果项目测试规模较小，可以全部映射到同一个完整测试命令。
+本项目把测试分为三层，按场景选择运行命令；新增测试文件默认归入 **full**，确认足够快且足够核心后，再上调到 core 或 smoke。
 
 ### smoke（目标 <5s）
 
 ```bash
-# TODO: 替换为本项目的 smoke 子集命令
-# npm run test:smoke       (Node.js)
-# pytest -m "not slow"     (Python)
-# go test -short ./...     (Go)
+npm run test:smoke
 ```
 
 适用场景：
-- implement-task / refine-task 内循环
+- code-task 内循环
 - 保存即跑 / 频繁反馈
 - 仅断言项目结构、配置、模板契约
 
 ### core（目标 <15s）
 
 ```bash
-# TODO: 替换为本项目的 core 子集命令
-# npm run test:core        (Node.js)
-# pytest -m "not contract" (Python)
-# go test ./...            (Go)
+npm run test:core
 ```
 
 适用场景：
 - pre-commit hook（自动调用）
-- 写 implementation.md / refinement.md 报告前的最终验证
+- 写 code.md / code-r{N}.md 报告前的最终验证
 - 推送 PR 前的本地把关
 
-### full（完整测试）
+### full（目标 <60s）
 
 ```bash
-# TODO: 替换为本项目的完整测试命令
-# npm test                 (Node.js)
-# mvn test                 (Maven)
-# pytest                   (Python)
-# go test ./...            (Go)
+npm test
 ```
 
 适用场景：
 - release / tag 前
-- CI
+- CI（unit-tests.yml）
 - main 合并前的最终把关
 
-如果项目暂不分层，smoke / core / full 可以全部映射到同一个完整测试命令；分层是反馈速度优化项，不是使用协作工作流的前置条件。
+full 层运行全部项目测试。`npm test` 使用通配匹配项目测试文件，**新增的测试文件会自动归入 full**，这是安全网。
 
 ## 3. 输出结果
 

package/templates/.agents/skills/update-agent-infra/scripts/sync-templates.js

@@ -23,6 +23,7 @@ const DEFAULTS = {
   "platform": {
     "type": "github"
   },
+  "requiresPullRequest": true,
   "sandbox": {
     "engine": null,
     "runtimes": [

package/templates/.agents/templates/task.en.md

@@ -11,7 +11,7 @@ priority:                       # Optional Issue field: Urgent | High | Medium |
 effort:                         # Optional Issue field: High | Medium | Low
 start_date:                     # Optional Issue field for Feature: YYYY-MM-DD
 target_date:                    # Optional Issue field for Feature: YYYY-MM-DD
-current_step: analysis         # analysis | design | implementation | review | fix | commit
+current_step: requirement-analysis # requirement-analysis | requirement-analysis-review | technical-design | technical-design-review | code | code-review | completed
 assigned_to:                   # claude | codex | gemini | opencode | human
 ---
 
@@ -46,11 +46,11 @@ assigned_to:                   # claude | codex | gemini | opencode | human
 
 ## Implementation Notes
 
-[Notes from the implementation phase. Decisions made, trade-offs, deviations from design.]
+[Notes from the code phase. Decisions made, trade-offs, deviations from design.]
 
 ## Review Feedback
 
-<!-- Populated by review-task -->
+<!-- Populated by review-* -->
 
 ## Activity Log
 

package/templates/.agents/templates/task.zh-CN.md

@@ -11,7 +11,7 @@ priority:                       # 可选 Issue 字段：Urgent | High | Medium |
 effort:                         # 可选 Issue 字段：High | Medium | Low
 start_date:                     # Feature 可选 Issue 字段：YYYY-MM-DD
 target_date:                    # Feature 可选 Issue 字段：YYYY-MM-DD
-current_step: analysis         # analysis | design | implementation | review | fix | commit
+current_step: requirement-analysis # requirement-analysis | requirement-analysis-review | technical-design | technical-design-review | code | code-review | completed
 assigned_to:                   # claude | codex | gemini | opencode | human
 ---
 
@@ -50,7 +50,7 @@ assigned_to:                   # claude | codex | gemini | opencode | human
 
 ## 审查反馈
 
-<!-- 由 review-task 填写 -->
+<!-- 由 review-* 填写 -->
 
 ## 活动日志
 

package/templates/.agents/workflows/bug-fix.en.yaml

@@ -1,149 +1,195 @@
-# Bug Fix Workflow
-# Use this workflow when fixing a bug.
+# Bug-fix workflow
+# Use this workflow when fixing a defect.
+#
+# Note: a step's `pr_tasks` list is counted toward workflow progress only when
+# `.agents/.airc.json:requiresPullRequest !== false` (see .agents/skills/complete-task/SKILL.md).
 
 name: bug-fix
-description: Workflow for diagnosing and fixing a bug.
+description: Workflow for diagnosing and fixing a defect.
 
 steps:
   - name: analysis
-    description: Reproduce the bug, identify the root cause, and determine the scope of the fix.
+    description: Reproduce the defect, identify the root cause, and define the fix scope.
     recommended_agents:
       - claude
       - gemini
     tasks:
-      - Understand the bug report and expected behavior
-      - Reproduce the bug when possible
+      - Understand the defect report and expected behavior
+      - Reproduce the defect when possible
       - Trace the code path to identify the root cause
-      - Determine impacted files and components
-      - Record findings in the task file
+      - Determine affected files and components
+      - Record findings in the task workspace
     inputs:
-      - Bug report or issue description
+      - Defect report or Issue description
       - Reproduction steps
       - Project codebase
     artifact_versioning:
       outputs:
         - name: analysis
           pattern: "analysis.md | analysis-r{N}.md"
-          rule: "Scan existing analysis artifacts; create analysis.md for the first round, then analysis-r{N}.md for later revisions (N = current highest round + 1)"
+          rule: "Scan existing analysis artifacts; create analysis.md for round 1 and analysis-r{N}.md for later revisions (N = current highest round + 1)"
     outputs:
-      - Root cause analysis
-      - List of impacted files
-      - Updated task file (analysis section)
+      - Root-cause analysis
+      - Affected files list
+      - Updated task analysis notes
+
+  - name: analysis-review
+    description: Review whether the defect analysis has sufficient reproduction evidence, a clear root cause, and a scoped fix.
+    recommended_agents:
+      - claude
+      - gemini
+    tasks:
+      - Verify reproduction evidence or the non-reproducible explanation
+      - Check root-cause reasoning and impact scope
+      - Confirm the analysis is ready for design
+      - Create the requirement analysis review report
+    inputs:
+      - Latest root-cause analysis
+      - Defect report or Issue description
+    artifact_versioning:
+      inputs:
+        - name: analysis
+          pattern: "analysis.md | analysis-r{N}.md"
+          rule: "Read the highest-round analysis artifact as review input"
+      outputs:
+        - name: review-analysis
+          pattern: "review-analysis.md | review-analysis-r{N}.md"
+          rule: "Scan existing requirement-analysis review artifacts; create review-analysis.md for round 1 and review-analysis-r{N}.md for later rounds (N = current highest round + 1)"
+    outputs:
+      - Requirement analysis review report
+      - Findings to fix, if any
 
   - name: design
-    description: Plan the fix and consider edge cases.
+    description: Plan the fix and account for edge cases.
     recommended_agents:
       - claude
     tasks:
-      - Determine the best approach for fixing the bug
-      - Identify potential side effects of the fix
-      - Plan test cases to validate the fix
-      - Record the fix plan in the task file
+      - Determine the best fix approach
+      - Identify potential side effects
+      - Plan regression tests
+      - Record the fix plan in the task workspace
     inputs:
-      - Root cause analysis from the previous step
+      - Approved root-cause analysis
       - Project architecture
     artifact_versioning:
       inputs:
         - name: analysis
           pattern: "analysis.md | analysis-r{N}.md"
           rule: "Read the highest-round analysis artifact as design input"
+        - name: review-analysis
+          pattern: "review-analysis.md | review-analysis-r{N}.md"
+          rule: "Read the highest-round analysis review artifact and confirm feedback is approved or handled"
       outputs:
         - name: plan
           pattern: "plan.md | plan-r{N}.md"
-          rule: "Scan existing plan artifacts; create plan.md for the first round, then plan-r{N}.md for later revisions (N = current highest round + 1)"
+          rule: "Scan existing plan artifacts; create plan.md for round 1 and plan-r{N}.md for later revisions (N = current highest round + 1)"
     outputs:
-      - Fix plan recorded in the task file
-      - List of test cases to add
+      - Fix plan
+      - Regression test list
 
-  - name: implementation
-    description: Implement the bug fix and add regression tests.
+  - name: design-review
+    description: Review whether the fix plan addresses the root cause, has sufficient tests, and controls side effects.
     recommended_agents:
-      - codex
-      - cursor
+      - claude
+      - gemini
     tasks:
-      - Create a bug-fix branch
-      - Implement the fix
-      - Add regression tests that would fail without the fix
-      - Verify that existing tests still pass
-      - Ensure the code follows project conventions
+      - Verify the plan covers the approved root-cause analysis
+      - Check regression test strategy and potential side effects
+      - Confirm the plan is ready for coding
+      - Create the plan review report
     inputs:
-      - Fix plan from the previous step
-      - Project coding standards
+      - Latest fix plan
+      - Latest root-cause analysis and review report
     artifact_versioning:
+      inputs:
+        - name: plan
+          pattern: "plan.md | plan-r{N}.md"
+          rule: "Read the highest-round plan artifact as review input"
       outputs:
-        - name: implementation
-          pattern: "implementation.md | implementation-r{N}.md"
-          rule: "Scan existing implementation artifacts; create implementation.md for the first round, then implementation-r{N}.md for later re-implementations (N = current highest round + 1)"
+        - name: review-plan
+          pattern: "review-plan.md | review-plan-r{N}.md"
+          rule: "Scan existing plan review artifacts; create review-plan.md for round 1 and review-plan-r{N}.md for later rounds (N = current highest round + 1)"
     outputs:
-      - Bug-fix branch with the implementation
-      - Regression test files
-      - Updated task file (implementation notes)
+      - Plan review report
+      - Findings to fix, if any
 
-  - name: review
-    description: Verify that the fix is correct and complete.
+  - name: code
+    description: Implement the defect fix and regression coverage; run the same stage again after code review to handle fixes.
     recommended_agents:
-      - claude
+      - codex
+      - cursor
     tasks:
-      - Verify that the fix addresses the root cause
-      - Check whether regression tests are sufficient
-      - Ensure no new issues were introduced
-      - Verify that the fix does not break existing functionality
-      - Create a review report
+      - Create a bug-fix branch
+      - Implement the fix
+      - Add a regression test that fails without the fix
+      - Fix findings from the latest code-review artifact when in fix mode
+      - Verify existing tests still pass
+      - Ensure code follows project conventions
     inputs:
-      - Bug-fix branch
-      - Root cause analysis
+      - Approved fix plan
+      - Project coding standards
+      - Latest code review report when in fix mode
     artifact_versioning:
       inputs:
-        - name: implementation
-          pattern: "implementation.md | implementation-r{N}.md"
-          rule: "Read the highest-round implementation artifact as review input"
+        - name: plan
+          pattern: "plan.md | plan-r{N}.md"
+          rule: "Read the highest-round plan artifact as implementation input"
+        - name: review-plan
+          pattern: "review-plan.md | review-plan-r{N}.md"
+          rule: "Read the highest-round plan review artifact and confirm feedback is approved or handled"
+        - name: review-code
+          pattern: "review-code.md | review-code-r{N}.md"
+          rule: "In fix mode, read the highest-round code review artifact and verify it matches the latest Code Review entry in task.md Activity Log"
       outputs:
-        - name: review
-          pattern: "review.md | review-r{N}.md"
-          rule: "Scan existing review artifacts; create review.md for the first round, then review-r{N}.md for later rounds (N = current highest round + 1)"
+        - name: code
+          pattern: "code.md | code-r{N}.md"
+          rule: "Scan existing code artifacts; create code.md for round 1 and code-r{N}.md for later fixes or reimplementation (N = current highest round + 1)"
     outputs:
-      - Review report
-      - List of issues if any
+      - Bug-fix branch with implementation or fixes
+      - Regression test files
+      - Updated task notes
 
-  - name: fix
-    description: Address issues found during review.
+  - name: code-review
+    description: Verify that the fix is correct and complete.
     recommended_agents:
-      - codex
-      - cursor
+      - claude
     tasks:
-      - Fix issues identified in the review
-      - Update tests as needed
-      - Verify that all tests pass
+      - Verify the fix addresses the root cause
+      - Check regression test sufficiency
+      - Ensure no new problems were introduced
+      - Verify the fix does not break existing behavior
+      - Create the code review report
     inputs:
-      - Review report
       - Bug-fix branch
+      - Root-cause analysis
     artifact_versioning:
       inputs:
-        - name: review
-          pattern: "review.md | review-r{N}.md"
-          rule: "Read the highest-round review artifact and verify that it matches the latest Code Review entry in the task.md Activity Log"
+        - name: code
+          pattern: "code.md | code-r{N}.md"
+          rule: "Read the highest-round code artifact as review input"
       outputs:
-        - name: refinement
-          pattern: "refinement.md | refinement-r{N}.md"
-          rule: "Scan existing refinement artifacts; create refinement.md for the first round, then refinement-r{N}.md for later rounds (N = current highest round + 1)"
+        - name: review-code
+          pattern: "review-code.md | review-code-r{N}.md"
+          rule: "Scan existing code review artifacts; create review-code.md for round 1 and review-code-r{N}.md for later rounds (N = current highest round + 1)"
     outputs:
-      - Updated branch with fixes
-      - Updated task file
+      - Code review report
+      - Findings list, if any
 
   - name: commit
-    description: Finalize the bug fix and create a pull request.
+    description: Finalize the defect fix and create a pull request (the PR portion runs only when the project enables the PR flow).
     recommended_agents:
       - claude
       - human
     tasks:
       - Ensure all tests pass
-      - Write a commit message that references the bug or issue
-      - Create a pull request with a bug-fix summary
-      - Link the PR to the bug report issue
+      - Write a commit message that references the defect or Issue
       - Move the task to completed
+    pr_tasks:
+      - Create a pull request with a defect-fix description
+      - Link the PR to the defect report Issue
     inputs:
       - Final bug-fix branch
-      - Task file
+      - Task workspace
     outputs:
-      - Pull request
-      - Completed task file (stored in .agents/workspace/completed/)
+      - Completed task workspace under .agents/workspace/completed/
+      - Pull request (only when the project enables the PR flow)