coding-agent-helpers/compact-debug-ledger
Use when a debugging thread needs to be compressed into a reusable investigation ledger. Capture the target, evidence, attempted fixes, ruled-out hypotheses, viable hypotheses, and next experiments. Good triggers include "compact this debugging session", "summarize what we've tried", and "turn this into a debugging ledger".
99
100%
Does it follow best practices?
Impact
99%
3.66xAverage score across 8 eval scenarios
Passed
No known issues
Evaluation results
100%
74%Debugging Session Summary
Required output section structure
Debug Target section
0%
100%
Debug Target one sentence
0%
100%
Evidence section
0%
100%
Evidence as bullet facts
0%
100%
Attempts section
0%
100%
Attempt status labels
0%
100%
Ruled Out section
0%
100%
Still Plausible section
0%
100%
Next Experiments section
0%
100%
No off-topic detail
100%
100%
Evidence over chronology
100%
100%
Output saved to file
100%
100%
95%
86%Compressing a Node.js Memory Leak Investigation
Attempt outcome labeling
Attempts section present
0%
100%
All attempts labeled
0%
100%
Correct worked labels
0%
58%
Correct failed labels
0%
100%
Ruled Out section
60%
100%
Still Plausible section
0%
100%
Next Experiments section
0%
100%
Next experiments count
30%
100%
Evidence section
0%
100%
Debug Target section
0%
100%
100%
92%Preserving a Complex Intermittent Bug Investigation
Hypothesis separation ruled-out vs plausible
Ruled Out section present
0%
100%
Ruled Out correct content
0%
100%
Still Plausible section present
0%
100%
Still Plausible correct content
0%
100%
No hypothesis mixing
0%
100%
Debug Target section
0%
100%
Evidence section
0%
100%
Attempts section
0%
100%
Next Experiments present
0%
100%
Output saved to file
100%
100%
100%
73%Active Bug Hunt: Flaky CI Pipeline
Next experiments count constraint (1-3)
Next Experiments section present
0%
100%
Experiments count limit
0%
100%
Experiments reduce uncertainty
66%
100%
Debug Target section
0%
100%
Evidence section
0%
100%
Attempts section with labels
50%
100%
Ruled Out section
0%
100%
Still Plausible section
0%
100%
File created
100%
100%
No dead detail
87%
100%
100%
82%Compressing a Multi-System Debugging Session
Debug target as single sentence
Debug Target section present
0%
100%
Debug Target is one sentence
0%
100%
Debug Target captures essence
0%
100%
Evidence section
0%
100%
Attempts section
0%
100%
Ruled Out section
0%
100%
Still Plausible section
0%
100%
Next Experiments section
0%
100%
No chronological replay
100%
100%
File saved
100%
100%
100%
54%Restructuring a Day-Long Debug Log
Evidence and hypotheses over chronology
No timestamp structure
66%
100%
Evidence section present
0%
100%
Evidence captures key facts
60%
100%
No irrelevant detail
100%
100%
Attempts section with labels
0%
100%
Ruled Out section
50%
100%
Still Plausible section
0%
100%
Next Experiments section
50%
100%
Debug Target one sentence
0%
100%
File saved
100%
100%
99%
46%Compacting a Verbose Security Incident Investigation
Strip dead conversation detail
No catch-up repetition
100%
100%
No absence/return mentions
100%
100%
Evidence preserved
53%
100%
Attempts section with labels
0%
100%
Ruled Out section
0%
100%
Still Plausible section
40%
100%
Next Experiments section
25%
100%
Debug Target one sentence
0%
100%
No repeated context
71%
85%
File saved
100%
100%
100%
68%Compressing a Code-Heavy Debugging Session
Minimal implementation detail for next experiment
No full code blocks
0%
100%
No exhaustive call-site list
75%
100%
Key insight preserved
100%
100%
Attempts section with labels
0%
100%
Evidence section
0%
100%
Ruled Out section
0%
100%
Still Plausible section
0%
100%
Next Experiments section
37%
100%
Debug Target one sentence
0%
100%
File saved
100%
100%
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents