Content
82%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.
An unusually disciplined doctrine skill: decision tables replace prose, every rule is stated once with concrete repo exceptions, and references out to other units are clearly signaled at one level. The main gaps are the absence of worked examples showing the doctrine applied and the lack of an explicit ordered authoring sequence.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is relentlessly lean: decision tables over prose ('Decision tables > prose explanations'), one-term-per-concept discipline, and no explanation of anything Claude already knows. Every exception clause cites a concrete repo case that earns its tokens. | 5 / 5 |
Actionability | Concrete, specific doctrine throughout — the unit-choice table, 'Phrase design decisions as "X over Y: reason"', 'Size a skill by a deletion pass using a /reduce loop', and pointers to executable references like 'hope/hooks/judge.sh'. It falls short of fully copy-paste-ready because it offers no worked example of a well-formed description or hook. | 4 / 5 |
Workflow Clarity | The section order (Frontmatter, Sizing, Skill Design, Hook Design) tracks the authoring process and the three decision tables act as checkpoints for each judgment call. It is not a linear procedure with explicit validation loops, so it does not reach the top anchor. | 4 / 5 |
Progressive Disclosure | No bundle files exist and none are needed; the ~80-line body is organized under clear section headers with well-signaled one-level references to repo resources ('hope/hooks/judge.sh', 'plugin-authoring', '/plugin-types'). Content appropriately lives in a single file for a doctrine skill, with only minor room to split skill-design vs hook-design material. | 4 / 5 |
Total | 17 / 20 Passed |