Content
93%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 exceptionally tight, high-signal skill body: fully executable examples and commands, near-zero padding, and a well-signaled one-level-deep reference for the edge case. The only notable gap is workflow sequencing — the verify/serve/hand-off flow is implied by section order with a check-until-clean loop, but lacks explicit step ordering and error-recovery guidance.
Suggestions
Add an explicit ordered build-to-delivery sequence (author the island and scenes, port source widgets, run `npx hyperframes check --json`, fix reported violations and re-run, then `npx hyperframes present`) so the workflow reads as steps rather than sections.
Include brief error-recovery guidance for `check --json` failures, e.g. how to interpret and fix unresolved `sceneId`s, dangling hotspot `target`s, or out-of-range fragments before re-validating.
Surface the standalone-harness reference as a markdown link with a one-line summary of what it covers, so navigation to the edge-case material is even more discoverable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is ~58 lines of dense, unpadded rules where every sentence carries a non-obvious constraint (e.g. "Hidden stacked slides need `opacity: 0`, `visibility: hidden` and `pointer-events: none` together, or an invisible layer swallows clicks") with no explanations of concepts Claude already knows. | 5 / 5 |
Actionability | It provides a copy-paste-ready JSON island example covering the common cases (notes, fragments, hotspots, sequences) and concrete commands such as `npx hyperframes check --json`, `npx hyperframes present <project-dir>`, `snapshot --at` each slide, and the catalog search invocation. | 5 / 5 |
Workflow Clarity | The phases (author the deck, port the page, verify with "Run `npx hyperframes check --json` until it is clean", serve with `present`, hand off the URL) form a clear sequence with an explicit validation loop, but the steps are not explicitly ordered or numbered and there is no error-recovery guidance for what to do when check fails. This sits between anchors 4 and 5; the missing explicit step ordering and error-recovery feedback detail make 4 the best fit. | 4 / 5 |
Progressive Disclosure | The core authoring model stays inline while the standalone-harness edge case is clearly signaled and deferred one level deep ("A deck opened from the bare player bundle without `present` needs workarounds: `references/standalone-harness.md`"), and that referenced file exists in the bundle. | 5 / 5 |
Total | 19 / 20 Passed |