Content
81%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.
A highly actionable, well-sequenced authoring contract with executable schema, a worked example, and explicit validation feedback loops. Its main weakness is length and redundancy in the media/mute and standalone-interim sections, which dilute token efficiency and keep some inlinable detail in the main file.
Suggestions
Consolidate the repeated cross-realm iframe mute/pause rule into one canonical statement and reference it, instead of restating it in both Media cleanup and Global nav mute.
Extract the dense porting-source-pages rules into a references file (e.g. references/porting.md) and keep a concise summary plus pointer in the main body.
Trim the standalone 'interim' section's hedging about the not-yet-shipped engine-hosted path to a single sentence.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is information-dense and avoids explaining concepts Claude already knows, but at ~520 lines it carries notable redundancy — the iframe cross-realm/mute rule is restated across Media cleanup and Global nav mute, and the standalone 'interim' section hedges heavily ('durable answer... coming'). | 3 / 5 |
Actionability | Provides copy-paste-ready, executable guidance: the full JSON island schema, a complete worked HTML deck example, exact commands ('npx hyperframes lint', 'npx hyperframes present', 'bun run build'), and explicit lint rules covering the common cases. | 5 / 5 |
Workflow Clarity | Sequences the full authoring process (intent confirmation → scenes → island → slides → fragments → branching → validation) with explicit validation checkpoints and a fix-and-rerun feedback loop ('Fix the file and rerun until lint reports 0 error(s)'). | 5 / 5 |
Progressive Disclosure | One real, clearly signaled one-level-deep reference (skills/slideshow/references/standalone-harness.md) offloads the standalone bridge patterns, and the main body is well sectioned; however only a single reference exists and substantial detail (porting rules, media-cleanup, mute semantics) is inlined that could live in references. | 4 / 5 |
Total | 17 / 20 Passed |