Investigate query evaluation failures in the Knowledge Graph synthetic data pipeline. Use when queries fail or return unexpected results after running the evaluate binary.
81
76%
Does it follow best practices?
Impact
84%
1.10xAverage score across 3 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./.claude/skills/debug-clickhouse-queries/SKILL.mdDiagnose failing eval queries from evaluation report
UNKNOWN_COLUMN categorization
100%
100%
Column name mismatch diagnosis
100%
100%
Global sampling interpretation
12%
50%
Path-scoped vs global distinction
62%
100%
Sampling vs predicate hypothesis
75%
100%
Global count diagnostic SQL
87%
100%
Path entity count SQL
100%
100%
orbit query command usage
62%
100%
MERGED edge direction
25%
0%
DESCRIBE TABLE diagnostic
100%
100%
Prioritized fix list
100%
100%
Edge type verification SQL
87%
100%
Diagnose garbage enum values in synthetic data generation
Garbage enum identification
100%
100%
Root cause: fake_data.rs fallback
60%
100%
field.enum_values check
90%
100%
Both enum type variants
100%
100%
Ontology values identified
100%
100%
Verification SQL: state check
100%
100%
Verification SQL: user_type check
100%
100%
Ontology file location
50%
50%
No sampling issue misdiagnosis
100%
100%
Evidence from sample data
100%
100%
Diagnose edge direction and cardinality issues
Person-is-source convention
83%
100%
Query direction fix
100%
100%
AUTHORED per-target cardinality
40%
50%
AUTHORED per-target vs per-source diagnosis
41%
50%
AUTHORED cardinality fix
30%
50%
MERGED edge count assessment
0%
0%
Edge count verification SQL
75%
87%
Simulator.yaml fix specification
30%
50%
Query JSON fix specification
100%
100%
Entity count awareness
100%
100%
f5efc36
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.