nailedit confidence - no need to waste tokens on another assistant when it's a no brainer fix for Claude
This commit is contained in:
Fahad
104
tools/debug.py
104
tools/debug.py
@@ -50,7 +50,14 @@ DEBUG_INVESTIGATION_FIELD_DESCRIPTIONS = {
|
||||
"hypothesis": (
|
||||
"Formulate your current best guess about the underlying cause. This is a working theory and may evolve based on further evidence."
|
||||
),
|
||||
"confidence": "How confident you are in the current hypothesis: 'low', 'medium', or 'high'.",
|
||||
"confidence": (
|
||||
"How confident you are in the current hypothesis: "
|
||||
"'low' (initial theory), 'medium' (good evidence), 'high' (strong to very strong evidence), "
|
||||
"'nailedit' (ONLY use for final step and ONLY when you have found the EXACT root cause with 100% certainty AND "
|
||||
"identified a simple, minimal fix that requires no expert consultation. Use this ONLY "
|
||||
"for obvious bugs and logic errors that you ABSOLUTELY are certain about and have no doubts because you have"
|
||||
"successfully mapped out the code flow and the root cause behind the issue."
|
||||
),
|
||||
"backtrack_from_step": "If a previous step needs revision, specify the step number to backtrack from.",
|
||||
"continuation_id": "Continuation token used for linking multi-step investigations and continuing conversations after discovery.",
|
||||
"images": (
|
||||
@@ -197,7 +204,7 @@ class DebugIssueTool(BaseTool):
|
||||
},
|
||||
"confidence": {
|
||||
"type": "string",
|
||||
"enum": ["low", "medium", "high"],
|
||||
"enum": ["low", "medium", "high", "nailedit"],
|
||||
"description": DEBUG_INVESTIGATION_FIELD_DESCRIPTIONS["confidence"],
|
||||
},
|
||||
"backtrack_from_step": {
|
||||
@@ -344,43 +351,72 @@ class DebugIssueTool(BaseTool):
|
||||
if continuation_id:
|
||||
response_data["continuation_id"] = continuation_id
|
||||
|
||||
# If investigation is complete, call the AI model for expert analysis
|
||||
# If investigation is complete, decide whether to call expert analysis or proceed with minimal fix
|
||||
if not request.next_step_required:
|
||||
response_data["status"] = "calling_expert_analysis"
|
||||
response_data["investigation_complete"] = True
|
||||
|
||||
# Prepare consolidated investigation summary
|
||||
investigation_summary = self._prepare_investigation_summary()
|
||||
# Check if Claude has absolute certainty and can proceed with minimal fix
|
||||
if request.confidence == "nailedit":
|
||||
# Trust Claude's judgment completely - if it says nailedit, skip expert analysis
|
||||
response_data["status"] = "nailedit_confidence_proceed_with_fix"
|
||||
|
||||
# Call the AI model with full context
|
||||
expert_analysis = await self._call_expert_analysis(
|
||||
initial_issue=getattr(self, "initial_issue", request.step),
|
||||
investigation_summary=investigation_summary,
|
||||
relevant_files=list(self.consolidated_findings["relevant_files"]),
|
||||
relevant_methods=list(self.consolidated_findings["relevant_methods"]),
|
||||
final_hypothesis=request.hypothesis,
|
||||
error_context=self._extract_error_context(),
|
||||
images=list(set(self.consolidated_findings["images"])), # Unique images
|
||||
model_info=arguments.get("_model_context"), # Use pre-resolved model context from server.py
|
||||
arguments=arguments, # Pass arguments for model resolution
|
||||
request=request, # Pass request for model resolution
|
||||
)
|
||||
investigation_summary = self._prepare_investigation_summary()
|
||||
response_data["complete_investigation"] = {
|
||||
"initial_issue": getattr(self, "initial_issue", request.step),
|
||||
"steps_taken": len(self.investigation_history),
|
||||
"files_examined": list(self.consolidated_findings["files_checked"]),
|
||||
"relevant_files": list(self.consolidated_findings["relevant_files"]),
|
||||
"relevant_methods": list(self.consolidated_findings["relevant_methods"]),
|
||||
"investigation_summary": investigation_summary,
|
||||
"final_hypothesis": request.hypothesis,
|
||||
"confidence_level": "nailedit",
|
||||
}
|
||||
response_data["next_steps"] = (
|
||||
"Investigation complete with NAILED-IT confidence. You have identified the exact "
|
||||
"root cause and a minimal fix. Proceed directly with implementing the simple fix "
|
||||
"without requiring expert consultation. Focus on the precise, minimal change needed."
|
||||
)
|
||||
response_data["skip_expert_analysis"] = True
|
||||
response_data["expert_analysis"] = {
|
||||
"status": "skipped_due_to_nailedit_confidence",
|
||||
"reason": "Claude identified exact root cause with minimal fix requirement",
|
||||
}
|
||||
else:
|
||||
# Standard expert analysis for high/medium/low confidence
|
||||
response_data["status"] = "calling_expert_analysis"
|
||||
|
||||
# Combine investigation and expert analysis
|
||||
response_data["expert_analysis"] = expert_analysis
|
||||
response_data["complete_investigation"] = {
|
||||
"initial_issue": getattr(self, "initial_issue", request.step),
|
||||
"steps_taken": len(self.investigation_history),
|
||||
"files_examined": list(self.consolidated_findings["files_checked"]),
|
||||
"relevant_files": list(self.consolidated_findings["relevant_files"]),
|
||||
"relevant_methods": list(self.consolidated_findings["relevant_methods"]),
|
||||
"investigation_summary": investigation_summary,
|
||||
}
|
||||
response_data["next_steps"] = (
|
||||
"Investigation complete with expert analysis. Present the findings, hypotheses, "
|
||||
"and recommended fixes to the user. Focus on the most likely root cause and "
|
||||
"provide actionable implementation guidance."
|
||||
)
|
||||
# Prepare consolidated investigation summary
|
||||
investigation_summary = self._prepare_investigation_summary()
|
||||
|
||||
# Call the AI model with full context
|
||||
expert_analysis = await self._call_expert_analysis(
|
||||
initial_issue=getattr(self, "initial_issue", request.step),
|
||||
investigation_summary=investigation_summary,
|
||||
relevant_files=list(self.consolidated_findings["relevant_files"]),
|
||||
relevant_methods=list(self.consolidated_findings["relevant_methods"]),
|
||||
final_hypothesis=request.hypothesis,
|
||||
error_context=self._extract_error_context(),
|
||||
images=list(set(self.consolidated_findings["images"])), # Unique images
|
||||
model_info=arguments.get("_model_context"), # Use pre-resolved model context from server.py
|
||||
arguments=arguments, # Pass arguments for model resolution
|
||||
request=request, # Pass request for model resolution
|
||||
)
|
||||
|
||||
# Combine investigation and expert analysis
|
||||
response_data["expert_analysis"] = expert_analysis
|
||||
response_data["complete_investigation"] = {
|
||||
"initial_issue": getattr(self, "initial_issue", request.step),
|
||||
"steps_taken": len(self.investigation_history),
|
||||
"files_examined": list(self.consolidated_findings["files_checked"]),
|
||||
"relevant_files": list(self.consolidated_findings["relevant_files"]),
|
||||
"relevant_methods": list(self.consolidated_findings["relevant_methods"]),
|
||||
"investigation_summary": investigation_summary,
|
||||
}
|
||||
response_data["next_steps"] = (
|
||||
"Investigation complete with expert analysis. Present the findings, hypotheses, "
|
||||
"and recommended fixes to the user. Focus on the most likely root cause and "
|
||||
"provide actionable implementation guidance."
|
||||
)
|
||||
else:
|
||||
response_data["next_steps"] = (
|
||||
f"Continue investigation with step {request.step_number + 1}. "
|
||||
|
||||
Reference in New Issue
Block a user