Content
67%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 a solid, mostly executable runbook for driving Carto through the Membrane CLI, with explicit connection-state handling and a sensible proxy fallback. Its weaknesses are padding (marketing intro, hollow "Carto Overview" and "Popular actions" sections) and a broken "Step 2" reference that muddles the otherwise clear workflow sequence.
Suggestions
Cut the marketing intro paragraph and the empty "Carto Overview" bullet list ("Use action names and parameters as needed" adds no instruction) — Claude already knows what CARTO is.
Fix the dangling "skip to **Step 2**" reference by numbering the actual steps consistently (the poll section is only labeled "1b" and no Step 2 exists).
Either populate "Popular actions" with real action examples (e.g. listing datasets, running a SQL query against a connection) or remove the section, since it currently just repeats the action-search command from the previous section.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The bulk is efficient executable command documentation, but there is noticeable padding: the marketing-style CARTO intro paragraph ("Businesses across various industries use CARTO to understand spatial patterns"), an empty "Carto Overview" entity list followed only by "Use action names and parameters as needed", and a "Popular actions" section that contains no actions, just a repeat of the search command. | 3 / 5 |
Actionability | Concrete, executable bash commands throughout — CLI install, `membrane connection ensure "https://carto.com" --json`, `action run` with `--input` JSON, and a full proxy flags table. Minor gaps: the Dataset/Column/Connection/OAuth App list carries no actionable guidance and `--tenant` is shown without its value. | 4 / 5 |
Workflow Clarity | A clear sequence (install → login → ensure connection → poll state → resolve CLIENT_ACTION_REQUIRED → search actions → run → proxy fallback) with explicit state checkpoints and error recovery for CONFIGURATION_ERROR/SETUP_FAILED. Not 5 because "skip to **Step 2**" references a step that is never labeled (only a stray "1b" exists) and step numbering is inconsistent. | 4 / 5 |
Progressive Disclosure | A single-file skill (no references/, scripts/, or assets/ exist) with well-organized sections and nothing long enough to demand extraction into a separate file; the proxy flags table is compact and appropriately inline. Minor organization gaps (orphaned step numbering, an empty section) keep it below 5. | 4 / 5 |
Total | 15 / 20 Passed |