Compressed caveman-style prose for AI coding agents — cuts ~65% output tokens while keeping full technical accuracy
96
100%
Does it follow best practices?
Impact
96%
1.00xAverage score across 38 eval scenarios
Passed
No known issues
{
"context": "Tests whether the response describes a safe zero-downtime column rename strategy.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Uses expand-contract (multi-phase) approach",
"description": "Describes a phased approach: add new column, dual-write, backfill, switch reads, remove old column — NOT a direct ALTER TABLE RENAME",
"max_score": 15
},
{
"name": "Explains why direct rename is risky",
"description": "Notes that a direct column rename breaks running application code that references the old name",
"max_score": 10
},
{
"name": "Includes backfill step",
"description": "Describes backfilling existing data from old column to new column (in batches to avoid locking)",
"max_score": 10
},
{
"name": "Mentions dual-write period",
"description": "Application writes to both old and new columns during transition so no data is lost",
"max_score": 10
},
{
"name": "No incorrect information",
"description": "PostgreSQL DDL behavior and migration strategy are technically correct",
"max_score": 12
}
]
}evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
scenario-11
scenario-12
scenario-13
scenario-14
scenario-15
scenario-16
scenario-17
scenario-18
scenario-19
scenario-20
scenario-21
scenario-22
scenario-23
scenario-24
scenario-25
scenario-26
scenario-27
scenario-28
scenario-29
scenario-30
scenario-31
scenario-32
scenario-33
scenario-34
scenario-35
scenario-36
scenario-37