Content
71%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 highly actionable with executable examples and a clear generation workflow, but it is verbose from triple-language duplication of one endpoint and weak on progressive disclosure: per-gateway detail is inlined for HeyGen while other gateways reference non-existent Layer 3 files.
Suggestions
Collapse the curl/TypeScript/Python variants of the same HeyGen POST into one canonical example, or move the language-specific implementations into a separate reference file.
Create the referenced Layer 3 files (gemini-omni, seedance-2-0) under references/ or remove the dangling references, so per-gateway prompting guides are actually reachable.
Split the HeyGen-specific API reference (request fields, providers table, response formats) into its own reference file, leaving SKILL.md as a multi-gateway overview that points one level deep to each gateway.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The same HeyGen POST endpoint is implemented in curl, TypeScript, and Python, and the curl example is repeated again in 'Usage Examples', creating several padded/redundant sections that could be trimmed. | 3 / 5 |
Actionability | Provides fully executable, copy-paste-ready curl/TypeScript/Python with concrete endpoints, request fields, response shapes, and a polling implementation covering the common generation flow. | 5 / 5 |
Workflow Clarity | A clear numbered Default Workflow (submit, poll every 10s, return video_url) with status handling for completed/failed/not_found, though output validation is implicit rather than an explicit checkpoint. | 4 / 5 |
Progressive Disclosure | Section headers give some structure, but a full HeyGen API reference is inlined in a multi-gateway skill while 'Layer 3 gemini-omni'/'seedance-2-0' references point to files that do not exist in any bundle directory. | 3 / 5 |
Total | 15 / 20 Passed |