Content
71%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 exceptionally concise and cleanly organized with a clear sequenced workflow ending in an explicit sanity-check step. Its one real weakness is actionability: the guidance stays at the level of high-level hints with no concrete encoding rules, chart-selection guidance, or worked examples to execute against.
Suggestions
Add concrete encoding guidance, e.g. a short mapping of comparison type to chart type (comparison → bar, trend over time → line, part-to-whole → stacked bar) so 'Choose encodings' is executable.
Spell out the sanity-check step's failure handling — what to do when axes, units, or missing values look wrong (fix and re-check) — to add a feedback loop.
Include one compact worked example (data question → chosen encoding → rendered output with source note) so the process has a concrete model to follow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is a lean 20 lines with zero padding: every section ('When to use', 'Non-goals', 'Workflow') carries only instructions Claude does not already know. It fully matches anchor 5 — assumes Claude's competence and every token earns its place — and is clearly above the 'minor instances of over-explanation' level of anchor 4. | 5 / 5 |
Actionability | The workflow consists of high-level hints — 'Inspect the data and questions', 'Choose encodings for the comparison at hand', 'Produce the visual with source/assumption notes' — with no concrete guidance on which encodings suit which comparisons, no chart-type selection rules, and no examples. This matches anchor 2 (high-level hints missing the specific steps to execute); it is above anchor 1 because the steps do form a coherent directive process, and below anchor 3 because there are no concrete specifics or examples at all. | 2 / 5 |
Workflow Clarity | The four steps are clearly sequenced (inspect → choose encodings → produce → 'Sanity-check axes, units, and missing values'), and step 4 is an explicit validation checkpoint, matching anchor 4 (clear sequence, most checkpoints present, minor validation gaps). It is not anchor 5 because there is no feedback loop saying what to do when the sanity check fails, and not anchor 3 since the checkpoint is explicit rather than missing or implicit. | 4 / 5 |
Progressive Disclosure | Per the scoring guidance, a skill under 50 lines with no need for external references can score 5 with just well-organized sections, and this one has exactly that: 'When to use', 'Non-goals', and 'Workflow' are cleanly separated and there is nothing that belongs in a separate reference file. There are no references/ , scripts/, or assets/ directories, so nothing is being deferred or buried. | 5 / 5 |
Total | 16 / 20 Passed |