Content
57%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 content's main strength is actionability — five concrete, largely executable Python templates covering SQL read models, search indexing, and aggregation. Its weaknesses are the absence of any sequenced workflow with validation checkpoints for batch/DB operations, the all-in-one 490-line body with no progressive disclosure into reference files, and padding (ASCII diagram, concept table) that assumes too little of Claude.
Suggestions
Add a numbered implementation workflow (define projection → register → run → verify checkpoint/lag → rebuild) with an explicit validation step after applying events, since batch/DB operations currently lack any feedback loop.
Move the larger templates (Elasticsearch search projection, multi-table projection) into files under references/ and keep a compact base template plus do's/don'ts in SKILL.md, linked with clearly signaled one-level-deep references.
Trim the ASCII architecture diagram and projection-types table — Claude already knows CQRS/event-sourcing fundamentals — and fix the missing `import asyncio` and cross-template Event/Projection imports so templates are self-contained.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body opens with "Comprehensive guide to building projections" and spends tokens on a decorative ASCII architecture diagram and a projection-types table that restate CQRS/event-sourcing concepts Claude already knows, and the five ~70-line templates could be tightened — "mostly efficient but includes some unnecessary explanation", matching anchor 3 rather than anchor 2 since there is no tutorial-style padding prose. | 3 / 5 |
Actionability | The templates provide concrete, near-copy-paste asyncpg/Elasticsearch code with real SQL (e.g., the ON CONFLICT upsert in DailySalesProjection), but minor gaps remain: Template 1 calls "asyncio.sleep" without importing asyncio, and Templates 2-5 rely on Event/Projection defined only in Template 1, fitting anchor 4 rather than fully-executable anchor 5. | 4 / 5 |
Workflow Clarity | No sequenced implementation workflow is given (define projection → register → run → checkpoint → rebuild is only implicit in code), and despite batch event processing and database writes there is no validation or feedback loop — the "Don't skip error handling" bullet is advice, not a checkpoint — so per the rubric's cap for batch/DB operations without validation, workflow clarity cannot exceed 3. | 3 / 5 |
Progressive Disclosure | Section headers give the ~490-line body decent structure, but roughly 400 lines of templates are inlined in SKILL.md with no bundle files at all, matching anchor 3 ("content that should be separate is inline"); it is above anchor 2 because structure exists and navigation within the file is easy. | 3 / 5 |
Total | 13 / 20 Passed |