Skip to content

Commit 27f4310

Browse files
popup-kaypopup-kay
authored
Merge pull request #2 from popup-studio-ai/feature/bkit-gemini-v151-enhancement
Merging bkit-gemini v1.5.1 enhancement to main
2 parents ef7f43a + cb20592 commit 27f4310

File tree

122 files changed

+23686
-569
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

122 files changed

+23686
-569
lines changed

‎.gemini/context/agent-triggers.md‎

Lines changed: 22 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,22 @@
1+
## Agent Triggers (8 Languages)
2+
3+
Agents are auto-triggered by keywords in user messages:
4+
5+
| Keywords | Agent | Action |
6+
|----------|-------|--------|
7+
| verify, check, is this right?, 검증, 맞아?, 確認, 正しい?, 验证, 对吗? | `gap-detector` | Design-implementation gap analysis |
8+
| improve, fix, iterate, 개선, 고쳐, 改善, 直して, 改进, 修复 | `pdca-iterator` | Auto-improvement with Evaluator-Optimizer pattern |
9+
| analyze, quality, issues, 분석, 품질, 分析, 品質, 质量, 问题 | `code-analyzer` | Code quality and architecture check |
10+
| report, summary, status, 보고서, 요약, 報告, 概要, 报告, 总结 | `report-generator` | PDCA completion report |
11+
| help, beginner, how to, 도움, 초보, 助けて, 初心者, 帮助, 新手 | `starter-guide` | Beginner-friendly guidance |
12+
| validate design, spec check, 설계 검증, 設計検証, 设计验证 | `design-validator` | Design document completeness check |
13+
| QA, testing, docker logs, 테스트, 로그, テスト, ログ, 测试, 日志 | `qa-monitor` | Zero Script QA with log monitoring |
14+
| pipeline, where to start, 뭐부터, 어디서부터, 何から, 从哪里 | `pipeline-guide` | 9-phase development pipeline guide |
15+
| bkend, auth, login, database, 인증, 로그인, 認証, ログイン, 身份验证 | `bkend-expert` | bkend.ai BaaS integration |
16+
| microservices, k8s, architecture, 마이크로서비스, アーキテクチャ, 微服务 | `enterprise-expert` | Enterprise architecture decisions |
17+
| AWS, terraform, infrastructure, 인프라, インフラ, 基础设施 | `infra-architect` | Cloud infrastructure design |
18+
| team, project lead, CTO, 팀 구성, 팀장, チームリード, 团队领导 | `cto-lead` | CTO-level team orchestration |
19+
| UI, frontend, component, 프론트엔드, フロントエンド, 前端 | `frontend-architect` | Frontend architecture decisions |
20+
| security, vulnerability, OWASP, 보안, 취약점, セキュリティ, 安全 | `security-architect` | Security review and analysis |
21+
| requirements, feature spec, user story, 요구사항, 기능 정의, 要件定義, 需求 | `product-manager` | Requirements and feature specs |
22+
| test strategy, QA plan, quality, 테스트 전략, QA 계획, テスト戦略, 测试策略 | `qa-strategist` | Test strategy and quality planning |

‎.gemini/context/commands.md‎

Lines changed: 29 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,29 @@
1+
## Available Commands
2+
3+
### PDCA Commands (Primary)
4+
| Command | Description | Use When |
5+
|---------|-------------|----------|
6+
| `/pdca plan [feature]` | Create plan document | Starting new feature |
7+
| `/pdca design [feature]` | Create design document | Plan approved |
8+
| `/pdca do [feature]` | Implementation guide | Ready to code |
9+
| `/pdca analyze [feature]` | Run gap analysis | Implementation complete |
10+
| `/pdca iterate [feature]` | Auto-improvement loop | Gap < 90% |
11+
| `/pdca report [feature]` | Completion report | Gap >= 90% |
12+
| `/pdca status` | Check current status | Anytime |
13+
| `/pdca next` | Guide to next step | Unsure what's next |
14+
15+
### Level Commands (Project Initialization)
16+
| Command | Level | Stack |
17+
|---------|-------|-------|
18+
| `/starter` | Starter | HTML/CSS/JS, static sites |
19+
| `/dynamic` | Dynamic | Next.js, BaaS (bkend.ai) |
20+
| `/enterprise` | Enterprise | Microservices, K8s, Terraform |
21+
22+
### Utility Commands
23+
| Command | Description |
24+
|---------|-------------|
25+
| `/code-review` | Analyze code quality |
26+
| `/zero-script-qa` | Log-based QA testing |
27+
| `/development-pipeline` | 9-phase pipeline guide |
28+
| `/output-style` | Change output style |
29+
| `/bkit` | Show all bkit functions |

‎.gemini/context/feature-report.md‎

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
## bkit Feature Usage Report Format
2+
3+
**Required at the end of every response:**
4+
5+
```
6+
─────────────────────────────────────────────────
7+
📊 bkit Feature Usage
8+
─────────────────────────────────────────────────
9+
✅ Used: [Features used in this response]
10+
⏭️ Not Used: [Major unused features] (reason)
11+
💡 Recommended: [Features suitable for next task]
12+
─────────────────────────────────────────────────
13+
```
14+
15+
### Features to Report:
16+
- **PDCA Skill**: /pdca plan, design, do, analyze, iterate, report, status, next
17+
- **Agents**: gap-detector, pdca-iterator, code-analyzer, report-generator, starter-guide, design-validator, qa-monitor, pipeline-guide, bkend-expert, enterprise-expert, infra-architect, cto-lead, frontend-architect, security-architect, product-manager, qa-strategist
18+
- **Level Skills**: /starter, /dynamic, /enterprise
19+
- **Phase Skills**: /phase-1-schema ~ /phase-9-deployment
20+
- **Utility Skills**: /code-review, /zero-script-qa, /development-pipeline, /output-style

‎.gemini/context/pdca-rules.md‎

Lines changed: 32 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,32 @@
1+
## PDCA Workflow Rules
2+
3+
### Core Cycle
4+
1. **New feature request** → Check/create Plan document first (`/pdca plan`)
5+
2. **Plan complete** → Create Design document (`/pdca design`)
6+
3. **Design complete** → Start implementation (`/pdca do`)
7+
4. **After implementation** → Run Gap analysis (`/pdca analyze`)
8+
5. **Gap Analysis < 90%** → Auto-improvement iteration (`/pdca iterate`)
9+
6. **Gap Analysis >= 90%** → Generate completion report (`/pdca report`)
10+
11+
### Behavioral Guidelines
12+
- Always verify important decisions with the user - AI is not perfect
13+
- Prefer editing existing files over creating new ones
14+
- Follow existing code patterns and conventions
15+
- Include bkit Feature Usage report at the end of every response
16+
- Never skip PDCA phases - follow the order
17+
18+
### PDCA Phase Recommendations
19+
20+
| Current Status | Recommended Action |
21+
|----------------|-------------------|
22+
| No PDCA started | `/pdca plan {feature}` |
23+
| Plan completed | `/pdca design {feature}` |
24+
| Design completed | Start implementation or `/pdca do {feature}` |
25+
| Implementation done | `/pdca analyze {feature}` |
26+
| Gap < 90% | `/pdca iterate {feature}` |
27+
| Gap >= 90% | `/pdca report {feature}` |
28+
29+
### Automation Rules
30+
- Gap Analysis < 90% → Auto-improvement with pdca-iterator (max 5 iterations)
31+
- Gap Analysis >= 90% → Completion report with report-generator
32+
- New feature detected → Suggest starting PDCA with `/pdca plan`

‎.gemini/context/skill-triggers.md‎

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
## Skill Triggers (Auto-detection)
2+
3+
Skills are activated based on context and keywords:
4+
5+
| Keywords | Skill | Description |
6+
|----------|-------|-------------|
7+
| static site, portfolio, landing page, 정적 웹, 静的サイト, 静态网站 | `starter` | Static website development |
8+
| fullstack, login, signup, database, 풀스택, 로그인, フルスタック | `dynamic` | Fullstack with BaaS |
9+
| microservices, kubernetes, terraform, 마이크로서비스, 微服务 | `enterprise` | Enterprise architecture |
10+
| schema, data model, terminology, 스키마, 용어, スキーマ, 数据模型 | `phase-1-schema` | Data structure design |
11+
| convention, coding style, 컨벤션, コンベンション, 编码规范 | `phase-2-convention` | Coding standards |
12+
| mockup, prototype, wireframe, 목업, 프로토타입, モックアップ | `phase-3-mockup` | UI/UX prototyping |
13+
| API, REST, backend, endpoint, API 설계, API設計, API设计 | `phase-4-api` | Backend API design |
14+
| design system, components, tokens, 디자인 시스템, コンポーネント | `phase-5-design-system` | Component library |
15+
| UI implementation, state management, UI 구현, 状態管理 | `phase-6-ui-integration` | Frontend-backend integration |
16+
| SEO, security, meta tags, XSS, 보안, セキュリティ, 安全 | `phase-7-seo-security` | SEO and security hardening |
17+
| code review, architecture check, 코드 리뷰, コードレビュー | `phase-8-review` | Quality verification |
18+
| deployment, CI/CD, production, 배포, デプロイ, 部署 | `phase-9-deployment` | Production deployment |
19+
| mobile app, React Native, Flutter, 모바일 앱, モバイルアプリ | `mobile-app` | Mobile app development |
20+
| desktop app, Electron, Tauri, 데스크톱 앱, デスクトップアプリ | `desktop-app` | Desktop app development |

‎.gemini/context/tool-reference.md‎

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,19 @@
1+
## Tool Name Reference
2+
3+
bkit uses Gemini CLI native tool names:
4+
5+
| Tool | Purpose | Example |
6+
|------|---------|---------|
7+
| `write_file` | Create/overwrite files | Creating new components |
8+
| `replace` | Edit existing files | Modifying code |
9+
| `read_file` | Read file contents | Understanding code |
10+
| `read_many_files` | Read multiple files | Batch analysis |
11+
| `run_shell_command` | Execute shell commands | Git, npm, docker |
12+
| `glob` | Find files by pattern | `**/*.tsx` |
13+
| `grep` | Search file contents | Finding definitions |
14+
| `list_directory` | List directory contents | Exploring structure |
15+
| `web_search` | Search the web | Finding documentation |
16+
| `web_fetch` | Fetch URL content | Reading web pages |
17+
| `spawn_agent` | Launch sub-agent | Complex multi-step tasks |
18+
| `task_write` | Manage tasks | Task tracking |
19+
| `activate_skill` | Load skill context | On-demand expertise |

‎.gitignore‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -26,6 +26,7 @@ npm-debug.log*
2626

2727
# Local settings
2828
.claude/
29+
.bkit/
2930

3031
# PDCA artifacts
3132
docs/.pdca-snapshots/

‎CHANGELOG.md‎

Lines changed: 106 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,106 @@
1+
# Changelog
2+
3+
All notable changes to bkit-gemini will be documented in this file.
4+
5+
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
6+
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
7+
8+
## [1.5.1] - 2026-02-11
9+
10+
### Added
11+
12+
- **5 New Agents** (16 total): cto-lead, frontend-architect, security-architect, product-manager, qa-strategist
13+
- **3 New Hook Events** (10 total): BeforeModel, AfterModel, BeforeToolSelection
14+
- **7 Per-Skill Hook Scripts** + 2 utility modules in `hooks/scripts/skills/` and `hooks/scripts/utils/`
15+
- **4 Output Styles**: bkit-learning, bkit-pdca-guide, bkit-enterprise, bkit-pdca-enterprise
16+
- **Agent Memory System**: Per-agent persistent storage with project/user scope (`lib/core/agent-memory.js`, 214 lines)
17+
- **Context Hierarchy**: 4-level config merge - Plugin, User, Project, Session (`lib/context-hierarchy.js`, 209 lines)
18+
- **Skill Orchestrator**: Custom YAML parser, agent delegation, template loading (`lib/skill-orchestrator.js`, 708 lines)
19+
- **@import Modularization**: GEMINI.md split into 6 context modules in `.gemini/context/`
20+
- **Team Mode Foundation**: 3 new MCP tools - team_create, team_assign, team_status
21+
- **Enhanced TOML Commands**: `@{path}` file inclusion, `!{command}` shell execution, `{{args}}` interpolation
22+
- **Skill Metadata Extension**: 10+ YAML frontmatter fields for all 21 skills
23+
- **Agent Native Frontmatter**: model, tools, temperature, max_turns, timeout_mins for all 16 agents
24+
- **Comprehensive Test Suite**: 258 test cases across 10 categories, 63 automated core cases (100% pass rate)
25+
26+
### Enhanced
27+
28+
- **Permission Manager**: Glob pattern matching with PDCA phase restrictions in `before-tool.js`
29+
- **Context Fork**: LRU cache (limit 10), named snapshots, and diff capabilities in `context-fork.js`
30+
- **Import Resolver**: Variable substitution (`${variable}`), circular dependency detection, 5s cache
31+
- **SessionStart Hook**: Dynamic context injection with returning user detection, project level auto-detection (392 lines)
32+
- **BeforeAgent Hook**: 8-language intent detection with ambiguity scoring (186 lines)
33+
- **bkit.config.json**: 5 new configuration sections (outputStyles, agentMemory, team, contextHierarchy, skillOrchestrator)
34+
- **MCP Server**: Extended from 3 to 6 tools (753 lines)
35+
36+
## [1.5.0] - 2026-02-01
37+
38+
### Added
39+
40+
- Gemini CLI v0.28 compatibility verification (100% pass rate on all compatibility tests)
41+
- Platform Adapter architecture for clean Gemini CLI native integration
42+
- 11 specialized agents with Gemini native frontmatter
43+
- 21 domain skills with progressive disclosure
44+
- 7-event hook system: SessionStart, BeforeAgent, BeforeTool, AfterTool, AfterAgent, PreCompress, SessionEnd
45+
- MCP server with 3 tools: spawn_agent, list_agents, get_agent_info
46+
- Unified PDCA skill with 8 actions: plan, design, do, analyze, iterate, report, status, next
47+
- 8 TOML custom commands with advanced syntax support
48+
- Library core modules: permission.js, import-resolver.js, context-fork.js
49+
- bkit.config.json centralized configuration (7 sections)
50+
51+
## [1.4.0] - 2026-01-15
52+
53+
### Added
54+
55+
- 8-language auto-detection: English, Korean, Japanese, Chinese, Spanish, French, German, Italian
56+
- Implicit agent and skill trigger keywords for natural language activation
57+
- Ambiguity detection with automatic clarifying question generation
58+
- Automatic PDCA phase progression based on user actions
59+
- bkit Feature Usage Report system appended to every response
60+
61+
## [1.3.0] - 2026-01-01
62+
63+
### Added
64+
65+
- Task Management integration with PDCA workflow
66+
- Task Chain Auto-Creation for multi-step features
67+
- Evaluator-Optimizer pattern via pdca-iterator agent (from Anthropic's agent architecture)
68+
- Task ID persistence across sessions
69+
70+
## [1.2.0] - 2025-12-15
71+
72+
### Added
73+
74+
- Core Modularization: `lib/common.js` split into `lib/core/`, `lib/pdca/`, `lib/intent/`, `lib/task/`
75+
- Context Engineering framework with 7 library modules
76+
- Unified hook system architecture
77+
78+
## [1.1.0] - 2025-12-01
79+
80+
### Added
81+
82+
- 9-stage Development Pipeline (Schema through Deployment)
83+
- 3 Project Levels: Starter (static), Dynamic (fullstack), Enterprise (microservices)
84+
- Multilingual keyword detection foundation
85+
86+
## [1.0.0] - 2025-11-15
87+
88+
### Added
89+
90+
- Initial release as Gemini CLI extension
91+
- Fork from bkit-claude-code adapted for Gemini CLI compatibility
92+
- PDCA (Plan-Do-Check-Act) methodology implementation
93+
- Basic hook system for lifecycle event interception
94+
- Tool name mapping: Claude Code tools to Gemini CLI equivalents
95+
- GEMINI.md context file (converted from CLAUDE.md)
96+
- Apache 2.0 license
97+
98+
---
99+
100+
[1.5.1]: https://github.com/popup-studio-ai/bkit-gemini/compare/v1.5.0...v1.5.1
101+
[1.5.0]: https://github.com/popup-studio-ai/bkit-gemini/compare/v1.4.0...v1.5.0
102+
[1.4.0]: https://github.com/popup-studio-ai/bkit-gemini/compare/v1.3.0...v1.4.0
103+
[1.3.0]: https://github.com/popup-studio-ai/bkit-gemini/compare/v1.2.0...v1.3.0
104+
[1.2.0]: https://github.com/popup-studio-ai/bkit-gemini/compare/v1.1.0...v1.2.0
105+
[1.1.0]: https://github.com/popup-studio-ai/bkit-gemini/compare/v1.0.0...v1.1.0
106+
[1.0.0]: https://github.com/popup-studio-ai/bkit-gemini/releases/tag/v1.0.0

0 commit comments

Comments
 (0)