Content
87%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 lean and actionable, with copy-paste commands and good sectioning. Its one gap is the lack of explicit validation/retry steps for destructive and batch operations.
Suggestions
Add a validation checkpoint after destructive calls (e.g. confirm the page is trashed with `ntn pages get` before declaring success).
Show a validate→fix→retry loop for block children PATCH and workers deploy (e.g. inspect `ntn workers runs logs <run-id>` and redeploy on failure).
Note how to verify `ntn api ls`/`--spec` output before issuing raw mutations, so risky raw-API calls get a preflight check.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and command-first with no padding or explanations of concepts Claude already knows; time-sensitive material is isolated in a 'Version notes' section rather than inline. | 5 / 5 |
Actionability | Copy-paste-ready ntn and curl commands with concrete flags, parent-ref syntax, and input-syntax notes cover the common cases end to end. | 5 / 5 |
Workflow Clarity | Sections are well-organized but destructive/batch operations (pages trash, workers deploy, block children PATCH) lack explicit validate-then-retry checkpoints, capping workflow clarity at 3. | 3 / 5 |
Progressive Disclosure | Single SKILL.md with clear, well-organized section headers and no external bundle files — meets the simple-skill exception for full marks. | 5 / 5 |
Total | 18 / 20 Passed |