Content
92%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.
The body is highly actionable and well-structured with strong validation workflows and clean progressive disclosure to real reference files; the only minor weakness is a few inline deep-dive sections that slightly stretch conciseness.
Suggestions
Consider moving the extended QPE / conditional CALL / spatial code blocks into a reference file and keeping only a one-line pointer inline to tighten conciseness toward a 5.
The Syntax Traps inline table overlaps with references/syntax-traps.md; trimming the inline table to the top 5–6 traps and pointing out sooner would save tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense, terse tables and rule lists with no padding ('Label-free MATCH (n) forbidden unless bound'), but the ~390-line body inlines several deep sections (QPE, conditional CALL, spatial, dynamic labels) that could be more aggressively offloaded to references; a few minor trims possible. | 4 / 5 |
Actionability | Copy-paste-ready executable Cypher blocks, exact Query API v2 curl with real flags, specific EXPLAIN/PROFILE red-flag fixes, and concrete 'Wrong | Right' trap tables covering common cases. | 5 / 5 |
Workflow Clarity | Clear sequenced workflows with validation checkpoints: Pre-flight decision table, Schema-First Protocol priority order, Write Gate (EXPLAIN -> report -> user confirm), and a closing Checklist; batch/destructive CALL IN TRANSACTIONS ops carry explicit validation and a Failure Recovery feedback section. | 5 / 5 |
Progressive Disclosure | Clear overview body with one-level-deep, well-signaled references (indexes, cypher-syntax, syntax-traps, performance, advanced-patterns, apoc, graph-type) — all verified real files under references/ — each with a one-line 'Load on demand' description for easy navigation. | 5 / 5 |
Total | 19 / 20 Passed |