fix: removed use_websearch; this parameter was confusing Codex. It started using this to prompt the external model to perform searches! web-search is enabled by Claude / Codex etc by default and the external agent can ask claude to search on its behalf.
This commit is contained in:
Fahad
@@ -165,7 +165,7 @@ All tools that work with files support **both individual files and entire direct
|
||||
- `analysis_type`: architecture|performance|security|quality|general
|
||||
- `output_format`: summary|detailed|actionable
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for documentation and best practices - allows model to request Claude perform searches (default: true)
|
||||
- **Web search capability**: The assistant now automatically requests web searches when it needs current documentation or best practices—no parameter required
|
||||
|
||||
```
|
||||
"Analyze the src/ directory for architectural patterns" (auto mode picks best model)
|
||||
@@ -198,7 +198,7 @@ All tools that work with files support **both individual files and entire direct
|
||||
- `runtime_info`: Environment details
|
||||
- `previous_attempts`: What you've tried
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for error messages and solutions - allows model to request Claude perform searches (default: true)
|
||||
- **Web search capability**: Automatically initiates searches for relevant error messages or recent fixes when needed
|
||||
|
||||
```
|
||||
"Debug this logic error with context from backend/" (auto mode picks best model)
|
||||
@@ -213,7 +213,7 @@ All tools that work with files support **both individual files and entire direct
|
||||
- `focus_areas`: Specific aspects to focus on
|
||||
- `files`: Files or directories for context
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: max, Gemini only)
|
||||
- `use_websearch`: Enable web search for documentation and insights - allows model to request Claude perform searches (default: true)
|
||||
- **Web search capability**: Automatically calls for research when architecture references or external insights are required
|
||||
|
||||
```
|
||||
"Think deeper about my design with reference to src/models/" (auto mode picks best model)
|
||||
@@ -444,7 +444,7 @@ Claude can then search for these specific topics and provide you with the most c
|
||||
**Web search control:**
|
||||
Web search is enabled by default, allowing models to request Claude perform searches for current documentation and solutions. If you prefer the model to work only with its training data, you can disable web search:
|
||||
```
|
||||
"Use gemini to review this code with use_websearch false"
|
||||
"Use gemini to review this code and confirm whether any new framework changes affect the recommendation"
|
||||
```
|
||||
|
||||
## System Prompts
|
||||
@@ -467,4 +467,4 @@ Each tool has a unique system prompt that defines its role and approach:
|
||||
To modify tool behavior, you can:
|
||||
1. Edit prompts in `prompts/tool_prompts.py` for global changes
|
||||
2. Override `get_system_prompt()` in a tool class for tool-specific changes
|
||||
3. Use the `temperature` parameter to adjust response style (0.2 for focused, 0.7 for creative)
|
||||
3. Use the `temperature` parameter to adjust response style (0.2 for focused, 0.7 for creative)
|
||||
|
||||
@@ -45,7 +45,7 @@ This workflow ensures methodical analysis before expert insights, resulting in d
|
||||
- **Cross-file relationship mapping**: Understand dependencies and interactions
|
||||
- **Architecture visualization**: Describe system structure and component relationships
|
||||
- **Image support**: Analyze architecture diagrams, UML charts, flowcharts: `"Analyze this system diagram with gemini to understand the data flow and identify bottlenecks"`
|
||||
- **Web search capability**: When enabled with `use_websearch` (default: true), the model can request Claude to perform web searches and share results back to enhance analysis with current documentation, design patterns, and best practices
|
||||
- **Web search capability**: Automatically requests Claude to perform web searches when fresh documentation, patterns, or best practices are needed, ensuring the analysis stays current
|
||||
|
||||
## Tool Parameters
|
||||
|
||||
@@ -70,7 +70,6 @@ This workflow ensures methodical analysis before expert insights, resulting in d
|
||||
- `output_format`: summary|detailed|actionable (default: detailed)
|
||||
- `temperature`: Temperature for analysis (0-1, default 0.2)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for documentation and best practices (default: true)
|
||||
- `use_assistant_model`: Whether to use expert analysis phase (default: true, set to false to use Claude only)
|
||||
- `continuation_id`: Continue previous analysis sessions
|
||||
|
||||
@@ -196,4 +195,4 @@ After analysis: "Recommended searches for Claude: 'FastAPI async best practices
|
||||
- **Use `codereview`** for: Finding bugs and security issues with actionable fixes
|
||||
- **Use `debug`** for: Diagnosing specific runtime errors or performance problems
|
||||
- **Use `refactor`** for: Getting specific refactoring recommendations and implementation plans
|
||||
- **Use `chat`** for: Open-ended discussions about code without structured analysis
|
||||
- **Use `chat`** for: Open-ended discussions about code without structured analysis
|
||||
|
||||
@@ -28,7 +28,7 @@ and then debate with the other models to give me a final verdict
|
||||
- **File reference support**: `"Use gemini to explain this algorithm with context from algorithm.py"`
|
||||
- **Image support**: Include screenshots, diagrams, UI mockups for visual analysis: `"Chat with gemini about this error dialog screenshot to understand the user experience issue"`
|
||||
- **Dynamic collaboration**: Gemini can request additional files or context during the conversation if needed for a more thorough response
|
||||
- **Web search capability**: Analyzes when web searches would be helpful and recommends specific searches for Claude to perform, ensuring access to current documentation and best practices
|
||||
- **Web search awareness**: Automatically identifies when online research would help and instructs Claude to perform targeted searches using continuation IDs
|
||||
|
||||
## Tool Parameters
|
||||
|
||||
@@ -38,7 +38,6 @@ and then debate with the other models to give me a final verdict
|
||||
- `images`: Optional images for visual context (absolute paths)
|
||||
- `temperature`: Response creativity (0-1, default 0.5)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for documentation and insights (default: true)
|
||||
- `continuation_id`: Continue previous conversations
|
||||
|
||||
## Usage Examples
|
||||
@@ -74,11 +73,11 @@ and then debate with the other models to give me a final verdict
|
||||
- **Ask for trade-offs**: Request pros/cons for better decision-making
|
||||
- **Use conversation continuation**: Build on previous discussions with `continuation_id`
|
||||
- **Leverage visual context**: Include diagrams, mockups, or screenshots when discussing UI/UX
|
||||
- **Request web searches**: Ask for current best practices or recent developments in technologies
|
||||
- **Encourage research**: When you suspect documentation has changed, explicitly ask the assistant to confirm by requesting a web search
|
||||
|
||||
## When to Use Chat vs Other Tools
|
||||
|
||||
- **Use `chat`** for: Open-ended discussions, brainstorming, getting second opinions, technology comparisons
|
||||
- **Use `thinkdeep`** for: Extending specific analysis, challenging assumptions, deeper reasoning
|
||||
- **Use `analyze`** for: Understanding existing code structure and patterns
|
||||
- **Use `debug`** for: Specific error diagnosis and troubleshooting
|
||||
- **Use `debug`** for: Specific error diagnosis and troubleshooting
|
||||
|
||||
@@ -87,7 +87,6 @@ The above prompt will simultaneously run two separate `codereview` tools with tw
|
||||
- `severity_filter`: critical|high|medium|low|all (default: all)
|
||||
- `temperature`: Temperature for consistency (0-1, default 0.2)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for best practices and documentation (default: true)
|
||||
- `use_assistant_model`: Whether to use expert analysis phase (default: true, set to false to use Claude only)
|
||||
- `continuation_id`: Continue previous review discussions
|
||||
|
||||
@@ -171,4 +170,4 @@ Reviews include:
|
||||
- **Use `codereview`** for: Finding bugs, security issues, performance problems, code quality assessment
|
||||
- **Use `analyze`** for: Understanding code structure without finding issues
|
||||
- **Use `debug`** for: Diagnosing specific runtime errors or exceptions
|
||||
- **Use `refactor`** for: Identifying structural improvements and modernization opportunities
|
||||
- **Use `refactor`** for: Identifying structural improvements and modernization opportunities
|
||||
|
||||
@@ -69,7 +69,6 @@ Get a consensus from gemini supporting the idea for implementing X, grok opposin
|
||||
- `focus_areas`: Specific aspects to emphasize
|
||||
- `temperature`: Control consistency (default: 0.2 for stable consensus)
|
||||
- `thinking_mode`: Analysis depth (minimal/low/medium/high/max)
|
||||
- `use_websearch`: Enable research for enhanced analysis (default: true)
|
||||
- `continuation_id`: Continue previous consensus discussions
|
||||
|
||||
## Model Configuration Examples
|
||||
@@ -142,4 +141,4 @@ The consensus tool includes built-in ethical safeguards:
|
||||
- **Use `consensus`** for: Multi-perspective analysis, structured debates, major technical decisions
|
||||
- **Use `chat`** for: Open-ended discussions and brainstorming
|
||||
- **Use `thinkdeep`** for: Extending specific analysis with deeper reasoning
|
||||
- **Use `analyze`** for: Understanding existing systems without debate
|
||||
- **Use `analyze`** for: Understanding existing systems without debate
|
||||
|
||||
@@ -75,7 +75,6 @@ This structured approach ensures Claude performs methodical groundwork before ex
|
||||
**Model Selection:**
|
||||
- `model`: auto|pro|flash|flash-2.0|flashlite|o3|o3-mini|o4-mini|gpt4.1|gpt5|gpt5-mini|gpt5-nano (default: server default)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for documentation and solutions (default: true)
|
||||
- `use_assistant_model`: Whether to use expert analysis phase (default: true, set to false to use Claude only)
|
||||
|
||||
## Usage Examples
|
||||
@@ -208,4 +207,4 @@ After analysis: "Recommended searches for Claude: 'Django 4.2 migration error sp
|
||||
|
||||
**Step 3:** "Found suspicious async/await pattern in session_manager.py lines 45-67. The await might be missing exception handling. This could explain silent failures."
|
||||
|
||||
**Completion:** Investigation reveals likely root cause in exception handling, ready for expert analysis with full context.
|
||||
**Completion:** Investigation reveals likely root cause in exception handling, ready for expert analysis with full context.
|
||||
|
||||
@@ -149,7 +149,6 @@ Use zen and perform a thorough precommit ensuring there aren't any new regressio
|
||||
- `focus_on`: Specific aspects to focus on
|
||||
- `temperature`: Temperature for response (default: 0.2)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for best practices (default: true)
|
||||
- `use_assistant_model`: Whether to use expert validation phase (default: true, set to false to use Claude only)
|
||||
- `continuation_id`: Continue previous validation discussions
|
||||
|
||||
@@ -250,4 +249,4 @@ Validation results include:
|
||||
- **Use `precommit`** for: Validating changes before git commit, ensuring requirement compliance
|
||||
- **Use `codereview`** for: General code quality assessment without git context
|
||||
- **Use `debug`** for: Diagnosing specific runtime issues
|
||||
- **Use `analyze`** for: Understanding existing code without validation context
|
||||
- **Use `analyze`** for: Understanding existing code without validation context
|
||||
|
||||
@@ -94,7 +94,6 @@ security remediation plan using planner
|
||||
- `severity_filter`: critical|high|medium|low|all (default: all)
|
||||
- `temperature`: Temperature for analytical consistency (0-1, default 0.2)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: medium, Gemini only)
|
||||
- `use_websearch`: Enable web search for security best practices and vulnerability databases (default: true)
|
||||
- `use_assistant_model`: Whether to use expert security analysis phase (default: true)
|
||||
- `continuation_id`: Continue previous security audit discussions
|
||||
|
||||
@@ -219,4 +218,4 @@ Security audits include:
|
||||
- **Use `codereview`** for: General code quality with some security considerations
|
||||
- **Use `analyze`** for: Understanding security architecture without vulnerability assessment
|
||||
- **Use `debug`** for: Investigating specific security incidents or exploit attempts
|
||||
- **Use `precommit`** for: Pre-deployment security validation and change impact assessment
|
||||
- **Use `precommit`** for: Pre-deployment security validation and change impact assessment
|
||||
|
||||
@@ -25,7 +25,7 @@ with the best architecture for my project
|
||||
- **File reference support**: `"Use gemini to think deeper about my API design with reference to api/routes.py"`
|
||||
- **Image support**: Analyze architectural diagrams, flowcharts, design mockups: `"Think deeper about this system architecture diagram with gemini pro using max thinking mode"`
|
||||
- **Enhanced Critical Evaluation (v2.10.0)**: After Gemini's analysis, Claude is prompted to critically evaluate the suggestions, consider context and constraints, identify risks, and synthesize a final recommendation - ensuring a balanced, well-considered solution
|
||||
- **Web search capability**: When enabled (default: true), identifies areas where current documentation or community solutions would strengthen the analysis and suggests specific searches for Claude
|
||||
- **Web search capability**: Automatically identifies areas where current documentation or community solutions would strengthen the analysis and instructs Claude to perform targeted searches
|
||||
|
||||
## Tool Parameters
|
||||
|
||||
@@ -37,7 +37,6 @@ with the best architecture for my project
|
||||
- `images`: Optional images for visual analysis (absolute paths)
|
||||
- `temperature`: Temperature for creative thinking (0-1, default 0.7)
|
||||
- `thinking_mode`: minimal|low|medium|high|max (default: high, Gemini only)
|
||||
- `use_websearch`: Enable web search for documentation and insights (default: true)
|
||||
- `continuation_id`: Continue previous conversations
|
||||
|
||||
## Usage Examples
|
||||
@@ -94,4 +93,4 @@ This ensures you get both deep reasoning and practical, context-aware advice.
|
||||
- **Use `thinkdeep`** for: Extending specific analysis, challenging assumptions, architectural decisions
|
||||
- **Use `chat`** for: Open-ended brainstorming and general discussions
|
||||
- **Use `analyze`** for: Understanding existing code without extending analysis
|
||||
- **Use `codereview`** for: Finding specific bugs and security issues
|
||||
- **Use `codereview`** for: Finding specific bugs and security issues
|
||||
|
||||
Reference in New Issue
Block a user