This document lists all the source files examined during the creation of the comprehensive example projects analysis.
cli_agents/qwen-code/README.md- Main project overviewcli_agents/qwen-code/CHANGELOG.md- Version historycli_agents/qwen-code/docs/cli/authentication.md- Auth detailscli_agents/qwen-code/docs/cli/commands.md- Command referencecli_agents/qwen-code/docs/cli/index.md- CLI guidecli_agents/qwen-code/docs/tools/- Tools documentation
cli_agents/qwen-code/package.json- Dependencies and scriptscli_agents/qwen-code/packages/cli/src/ui/models/availableModels.ts- Model definitionscli_agents/qwen-code/packages/core/src/core/openaiContentGenerator/provider/dashscope.ts- DashScope provider implementationcli_agents/qwen-code/packages/core/src/core/openaiContentGenerator/provider/- Other provider implementations
- Monorepo architecture with core and CLI packages
- Vision model auto-switching feature
- DashScope cache control system
- Session token management
- Multi-authentication support (OAuth + API Key)
cli_agents/gemini-cli/README.md- Main project overviewcli_agents/gemini-cli/GEMINI.md- Development guidelinescli_agents/gemini-cli/docs/get-started/authentication.md- Auth setupcli_agents/gemini-cli/docs/cli/checkpointing.md- Session savingcli_agents/gemini-cli/docs/cli/token-caching.md- Token optimizationcli_agents/gemini-cli/docs/core/policy-engine.md- Policy managementcli_agents/gemini-cli/docs/tools/mcp-server.md- MCP integration
cli_agents/gemini-cli/package.json- Project structure and dependenciescli_agents/gemini-cli/packages/core/src/config/models.ts- Model definitions and fallbackcli_agents/gemini-cli/packages/core/src/core/client.ts- GeminiClient implementationcli_agents/gemini-cli/packages/core/src/core/geminiRequest.ts- Request buildingcli_agents/gemini-cli/packages/core/src/core/baseLlmClient.ts- Base LLM clientcli_agents/gemini-cli/packages/core/src/code_assist/- OAuth and credential handlingcli_agents/gemini-cli/packages/core/src/mcp/- MCP provider implementations
- Monorepo with core, CLI, and a2a-server packages
- Three authentication methods (OAuth, API Key, Vertex AI)
- Thinking mode with token limits
- Model fallback logic for service degradation
- Loop detection and chat compression services
- Comprehensive IDE context tracking
- Three-tier release management (nightly, preview, stable)
cli_agents/deepseek-cli/README.md- Main project overviewcli_agents/deepseek-cli/LOCAL-SETUP.md- Local setup guide
cli_agents/deepseek-cli/package.json- Minimal dependenciescli_agents/deepseek-cli/src/api.ts- Dual-mode API clientcli_agents/deepseek-cli/src/cli.ts- CLI interfacecli_agents/deepseek-cli/src/config.ts- Configuration managementcli_agents/deepseek-cli/src/commands/interactive.ts- Interactive REPLcli_agents/deepseek-cli/src/commands/chat.ts- Chat commandcli_agents/deepseek-cli/src/commands/setup.ts- Setup automationcli_agents/deepseek-cli/src/utils/exec.ts- Execution utilities
- Single-package simple architecture
- Dual local/cloud execution modes
- Ollama integration
- Minimal dependency footprint (5 core)
- Setup automation and health checking
- Response formatting with markdown support
- Model availability detection
- Examined README files for high-level overview
- Reviewed package.json for dependencies and structure
- Analyzed key source files for implementation patterns
- Reviewed documentation for features and capabilities
- Identified unique features and best practices
- Compared approaches across three projects
EXAMPLE_PROJECTS_ANALYSIS.md- Comprehensive 680-line analysisEXAMPLE_PROJECTS_QUICK_REFERENCE.md- Quick lookup guideANALYSIS_SOURCES.md- This file
- Image detection in prompts
- Vision model switching modes (once, session, persist)
- Vision-specific API parameters
- High-resolution image support
- Google Search grounding
- Search-aware response generation
- Integration with Gemini models
- Qwen: Session token limits and compression
- Gemini: Full checkpoint save/restore
- DeepSeek: Basic session history
- Qwen: OAuth + OpenAI-compatible fallback
- Gemini: OAuth → API Key → Vertex AI
- DeepSeek: Local Ollama → Cloud API
- Token caching (ephemeral)
- Chat compression
- Context window management
- Model fallback strategies
- MCP Protocol (Qwen, Gemini)
- File system operations
- Shell execution
- Web fetching and search
- Qwen DashScope:
Qwen_Code/packages/core/src/core/openaiContentGenerator/provider/dashscope.ts - Gemini Client:
Gemini_CLI/packages/core/src/core/client.ts - DeepSeek API:
DeepSeek_CLI/src/api.ts
- Qwen Auth:
Qwen_Code/packages/core/src/core/(OAuth & credential storage) - Gemini Auth:
Gemini_CLI/packages/core/src/code_assist/andpackages/core/src/core/ - DeepSeek Config:
DeepSeek_CLI/src/config.ts
- Qwen Models:
Qwen_Code/packages/cli/src/ui/models/availableModels.ts - Gemini Models:
Gemini_CLI/packages/core/src/config/models.ts - DeepSeek Models:
DeepSeek_CLI/src/config.ts
- Vision Switching:
Qwen_Code/packages/cli/src/ui/commands/modelCommand.ts - Checkpointing:
Gemini_CLI/packages/core/src/(ChatCompressionService) - Setup Automation:
DeepSeek_CLI/src/commands/setup.ts
- Import DeepSeek provider pattern (simplest)
- Add Qwen OAuth authentication
- Implement Gemini API key support
- Implement vision model auto-switching
- Add token caching strategies
- Implement model fallback logic
- Add MCP protocol support
- Integrate Google Search grounding
- Implement checkpointing system
- Add loop detection
- Implement chat compression
- Add IDE context tracking
- Analysis Date: November 6, 2025
- Projects Analyzed: 3
- Total Files Examined: 50+
- Documentation Generated: 3 markdown files
- Total Lines of Analysis: 1000+
All analysis documents are stored in /Users/milosvasic/Projects/helix_code/