You get:
- debugging from the wrong line (wasting time)
- missing the root cause buried in the trace
- no idea where to add logging or breakpoints
- ignoring the call chain that led to the error
- fixing symptoms instead of causes
But a stack trace is not random.
It is a map of the execution path.
- First frame: where the error was thrown (not always the cause)
- Your code frames: where your logic is (start here)
- Framework/library frames: rarely the problem
- Root cause: often several frames down
Without analysis, you fix the wrong line.
This framework forces AI to pinpoint the root cause.
Assume the role of a debugging expert who reads stack traces like a detective reads clues. Your task is to analyze a stack trace and identify the root cause. Generate: 1. ROOT CAUSE IDENTIFICATION - Which line/file is the actual cause - Why (not where) the error occurred 2. CALL CHAIN EXPLANATION - How execution got to the error - Which frames are relevant vs. irrelevant 3. LOGGING SUGGESTIONS - Where to add console.log or logging statements - What to log at each point 4. BREAKPOINT RECOMMENDATIONS - Which lines to set breakpoints - What to inspect at each breakpoint 5. FIX SUGGESTION - Specific code change INPUTS: Stack Trace (paste full trace): [PASTE STACK TRACE] Language/Framework: [PYTHON / JAVASCRIPT / JAVA / C# / OTHER] Known Recent Changes: [LIST OR "NONE"] Error Type (if known): [E.G., "Null reference," "Index out of range," "Type mismatch"] RULES: - Identify the first line of YOUR code in the stack (not library code) - Explain the call chain (how execution reached the error) - Ignore framework/library frames for root cause - Suggest specific logging (not "add more logs") - Suggest specific breakpoints (not "debug it") - Don't assume you know the fix without enough context
- Paste the full stack trace (not just the error message).
- Identify which lines are your code vs. library code.
- The root cause is usually in your code, not the library.
- Follow the logging suggestions to instrument your code.
- Set breakpoints at the recommended lines and inspect variables.
Stack Trace:
Traceback (most recent call last):
File “app.py”, line 45, in
result = process_data(user_input)
File “app.py”, line 32, in process_data
return [item.name for item in data_list]
File “app.py”, line 32, in
AttributeError: ‘NoneType’ object has no attribute ‘name’
Language/Framework: PYTHON
Known Recent Changes: Changed data_list to come from a new API endpoint instead of a local file
Error Type: AttributeError (‘NoneType’ object has no attribute ‘name’)
This framework improves outcomes by forcing:
- root cause identification (accuracy)
- call chain explanation (understanding)
- specific logging suggestions (instrumentation)
- breakpoint recommendations (debugging)
- fix suggestion (resolution)
Great stack trace analysis doesn’t just read — it interprets and prescribes.
Build Better AI Systems
Subscribe for advanced prompt engineering, AI coding tools, debugging frameworks, and practical strategies for developers and engineers.