Content
68%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.
Highly actionable content with strong copy-paste-ready examples and a reasonable reference structure, but it is verbose for an overview (citation boilerplate, known-concept explainers, repeated opener), lacks an end-to-end validated workflow, and leaves one bundle file (specialized-topics.md) orphaned from navigation.
Suggestions
Link specialized-topics.md from the 'Detailed Documentation' list (or remove the file) so every bundle file is reachable from the overview.
Move the inline 'Image Classification' and 'Modern Cloud-Native Workflows' sections into machine-learning.md and big-data.md respectively, keeping only a brief pointer in the body to improve progressive disclosure and conciseness.
Add an explicit validate-then-retry feedback loop to the batch classification example (e.g., sanity-check predicted class distribution / output raster before writing) so workflow_clarity can exceed the batch-operation cap of 3.
Trim the OGC-standards/CRS basics explainers and the citation-boilerplate section, and drop the repeated opening summary line, to tighten token efficiency.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly tight, actionable code, but it includes unnecessary material Claude already knows (the OGC-standards and basic-CRS explainers), a ~15-line citation-boilerplate section unrelated to the task, and an opening line that repeats the frontmatter description; the ~375-line body could also be tightened by pushing more inline examples into the existing reference files. | 3 / 5 |
Actionability | Packed with copy-paste-ready, executable Python for the common cases (NDVI from Sentinel-2, GeoPandas spatial join, GEE time series, RF classification, STAC+Planetary Computer load, COG read/write/validate) with real library calls and parameters, matching the fully-executable/common-cases 5 anchor. | 5 / 5 |
Workflow Clarity | The body is a topic catalog rather than a coherent sequenced workflow; scattered checkpoints exist (CRS assert, cog_validate, geometry is_valid) but the batch classification example predicts and writes output with no validation feedback loop, triggering the cap-at-3 rule for batch operations without validation. | 3 / 5 |
Progressive Disclosure | Good one-level-deep structure with a clearly labeled 'Detailed Documentation' section pointing to 13 reference files, but specialized-topics.md exists in references/ yet is never linked from the body (orphaned, unnavigable), and some heavy inline sections (Image Classification, Cloud-Native Workflows) overlap content that has dedicated reference files — minor organization gaps that keep it below 5. | 4 / 5 |
Total | 15 / 20 Passed |