Content
83%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 reference with lean prose and executable examples across SQL, pipelines, Cortex AI, Snowpark, and dbt. The main gap is the absence of explicit validation checkpoints for its batch/destructive operations, which caps workflow clarity.
Suggestions
Add explicit validation checkpoints to destructive/batch flows, e.g. 'After MERGE, verify row counts: SELECT count(*) before/after' or a validate-then-resume step for tasks.
Factor the Cortex AI function table and deprecated-function list into a references/cortex.md file and link from SKILL.md to reduce inline reference bulk.
Include a short 'Build a pipeline' numbered workflow (choose approach → create DT/stream → validate → resume task → monitor) with a verify step to lift workflow clarity above 3.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean — bullets, tables, and executable code with no padding and no explanation of concepts Claude already knows; every line earns its place (e.g. 'DataFrames are lazy — executed on collect()/show().'). | 5 / 5 |
Actionability | Provides copy-paste-ready, executable examples across domains — MERGE upserts, Dynamic Table DDL, Stream/Task DDL with RESUME, the Cortex Agent $spec$ block, Snowpark Session setup, and dbt config macros — covering the common cases. | 5 / 5 |
Workflow Clarity | Sequences are present (e.g. create stream → create task → ALTER TASK RESUME) and a Common Error Patterns table offers recovery, but batch/destructive operations (MERGE, pipelines, CREATE OR REPLACE) lack explicit validate-before-commit checkpoints, capping workflow clarity at 3 per the rubric. | 3 / 5 |
Progressive Disclosure | No bundle files exist; the skill is a single well-sectioned reference with clear headers and tables for navigation. Structure is good, though the Cortex function reference and code-pattern catalog could be split into separate reference files for a cleaner overview. | 4 / 5 |
Total | 17 / 20 Passed |