Content
78%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 well-engineered advisory skill body: highly actionable CLI guidance, strong destructive-operation safeguards, and an exemplary one-level-deep progressive disclosure structure backed by real bundle files. The main weakness is verbosity in the repeated per-scenario "Critical Facts" blocks, which inflate token cost without adding proportional signal.
Suggestions
Consolidate the five "Critical Facts to Always Surface" scenario blocks: factor the repeated patterns (engine detection via describe-db-instances, advisory-only stance, no-Data-API-for-MariaDB) into a shared "Critical Facts" reference and keep only scenario-unique facts inline to reduce token cost.
Tighten the Security Considerations list where it overlaps with the Production Instance Creation defaults (TLS, Secrets Manager, audit logging) to avoid restating the same guidance in two places.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient and assumes Claude's knowledge of RDS/AWS (it does not explain basic concepts), but the five "Critical Facts to Always Surface" blocks repeat similar guidance across scenarios and the RDS-vs-Aurora table plus the Security Considerations list restate several points, leaving room to tighten. | 3 / 5 |
Actionability | Provides fully executable, copy-paste-ready CLI commands with exact flags for the common cases — e.g. the full `aws rds create-db-instance` block and the parameter-group/TLS configuration steps — and names specific commands like `create-blue-green-deployment` and `switchover-blue-green-deployment` rather than generic descriptions. | 5 / 5 |
Workflow Clarity | Multi-step workflows are clearly sequenced with explicit validation checkpoints — prerequisite validation ("binlog_format=ROW, automated backups enabled"), a hard confirmation gate before the destructive switchover, and the "DO NOT bail out" rule that forces the full advisor workflow — satisfying the destructive-op validation requirement, with only minor gaps in some workflow references. | 4 / 5 |
Progressive Disclosure | SKILL.md is a clear overview that routes via a Decision Guide table to four well-signaled one-level-deep workflow references (upgrade, commitment-pricing, proxy, bluegreen), each of which in turn points to engine-specific sub-references that exist as real files in ./references; navigation is easy and nesting stays one level deep from each main workflow. | 5 / 5 |
Total | 17 / 20 Passed |