Content
75%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 well-structured, action-oriented body: concrete prompts, named models, aspect ratios, and clearly phased parallel generation with sensible conditional branches. The gaps are minor — no end-to-end CLI command examples, no validation of generated assets, and slight inline padding.
Suggestions
Show one complete example invocation (e.g. muapi image generate --model bytedance-seedream-v4.5 --prompt "..." --aspect-ratio 16:9) so the CLI argument syntax is unambiguous.
Add a light validation step after generation, e.g. re-check each image for consistent colors/typography against style_direction and regenerate any asset that drifted.
Trim the bold description restatement and the 'Trigger Keywords' section (or fold them into frontmatter) to tighten token usage.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean — a compact inputs table and dense prompt specs with almost no concept explanation Claude doesn't need — but has minor trimmable padding: the bold restatement of the description, the 'Trigger Keywords' section, and the awkward 'all in a single the plan' phrasing. Fits 'efficient; minor instances of over-explanation'. | 4 / 5 |
Actionability | Prompts are copy-paste ready with model IDs, aspect ratios, and a curl fallback, but the exact muapi CLI invocation syntax (how model, prompt, and aspect-ratio arguments are passed) is never shown, leaving minor gaps. Matches 'mostly executable guidance; concrete commands with minor gaps' — not anchor 5's fully copy-paste-ready command coverage. | 4 / 5 |
Workflow Clarity | The three parallel phases are clearly sequenced with numbered steps and explicit conditional handling ('If {{primary_color}} is given...', 'If {{existing_logo}} provided, use muapi image edit'), plus a polling hint — but there is no validation checkpoint that generated images match the brand direction before presenting them. Fits 'clear sequence with most checkpoints; minor validation gaps'. | 4 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are absent) and none are referenced, and the body is well-sectioned — but at ~62 lines it exceeds the under-50-line simple-skill threshold, and content like the trigger-keyword list and agent-operational notes arguably belong elsewhere or in frontmatter. 'Good structure; minor organization gaps' fits better than a clean 5. | 4 / 5 |
Total | 16 / 20 Passed |