Content
88%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A dense, highly actionable skill body with executable examples, explicit ordered checklists, and real validation feedback loops for destructive/batch operations. The main weakness is conciseness/structure: a large amount of reference material is inlined in SKILL.md rather than pushed to bundle files, slightly straining the context budget.
Suggestions
Move the full CALL IN TRANSACTIONS reference (syntax, REPORT STATUS columns, ON ERROR modes table) into a references/ file and keep only the load-bearing pattern inline in SKILL.md, linking out for the rest.
Extract the neo4j-admin header-file format tables and key-flags reference into a dedicated references/neo4j-admin-import.md, leaving the command skeleton and dry-run/validation steps in the main body.
Trim the version-number-heavy annotations (e.g. "2026.02+", "2025.12", "2026.06+") into a single "version compatibility" note or a deprecated/old-patterns subsection, as time-sensitive inline tags add tokens without aiding execution.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Largely lean — tables for ON ERROR modes, batch sizes, and error→fix mappings compress a lot of reference material without explaining concepts Claude already knows — but the ~500-line body inlines full reference content (e.g. the entire CALL IN TRANSACTIONS syntax and header-format tables) that could arguably live in a bundle file, leaving minor over-density. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready Cypher/bash examples for every method (LOAD CSV with type coercion, neo4j-admin import command, apoc.periodic.iterate), plus concrete config tables, header-file templates with sample data rows, and a Common Errors table mapping real error strings to fixes — covering the common cases comprehensively. | 5 / 5 |
Workflow Clarity | Explicit multi-step workflows with validation checkpoints throughout: an ordered Pre-Import Checklist ("Run in this exact order"), a destructive neo4j-admin offline load with a --dry-run validation step, and a Post-Import Validation section that polls populationPercent until 100 and gates production queries behind ONLINE indexes — exactly the validate→fix→retry feedback loop the rubric rewards. | 5 / 5 |
Progressive Disclosure | Well-signaled one-level-deep references to real bundle files ([driver-batch-write.md](references/driver-batch-write.md), [data-importer-gui.md](references/data-importer-gui.md), [post-import-refactoring.md](references/post-import-refactoring.md), confirmed present) plus external manual links, but the body itself carries substantial bulk reference material inline (full CALL IN TRANSACTIONS reference, header-format tables) that could be split out for cleaner separation. | 4 / 5 |
Total | 18 / 20 Passed |