Content
72%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, mostly executable recipe with concrete prompts, models, and a sensible approval-gated two-phase workflow. The main gaps are the lack of a validation/consistency-check step before batch-rendering six angles and CLI invocations that stop just short of full copy-paste commands.
Suggestions
Add an explicit validation checkpoint after Phase B, e.g. 'Before presenting the gallery, verify each output matches the approved base image (clothing, face, setting); regenerate any inconsistent angle with adjusted prompt wording.'
Provide complete CLI invocations with flags (e.g. 'muapi image generate --model nano-banana-pro --prompt "..." --aspect-ratio 3:4') so steps are copy-paste ready rather than implying the argument syntax.
Drop the verbatim bold restatement of the description under the title and fold the Trigger Keywords line into the frontmatter description to remove redundancy.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and operational: an inputs table, two phases, six ready-made prompts, and brief agent notes with no explanations of concepts Claude already knows. It falls short of 5 due to minor redundancy — the description is restated verbatim in bold under the title, and the Trigger Keywords section partially duplicates the description's angle vocabulary. | 4 / 5 |
Actionability | Concrete guidance throughout: named models ('nano-banana-pro', 'nano-banana-2-edit'), full copy-paste prompts per angle, aspect ratio guidance, and a fallback curl command with polling. It is not 5 because the exact muapi CLI flag syntax for generate/edit is implied ('muapi image generate (model=...)') rather than given as complete copy-paste-ready commands. | 4 / 5 |
Workflow Clarity | The two-phase sequence is clear with an explicit user-approval checkpoint ('Present the base image to the user for approval'), but Phase B submits all six angles as one batch with no validation or verification step (e.g., checking identity/clothing consistency before presenting the gallery). The batch-operation cap without validation limits this to 3 even though the sequence itself is well structured. | 3 / 5 |
Progressive Disclosure | The skill is a short, single-purpose recipe with no bundle files and no external references needed; sections (Inputs, Steps, Trigger Keywords, Notes) are well organized and everything is appropriately inline at one level. Per the simple-skill guideline, well-organized sections alone warrant the top score here. | 5 / 5 |
Total | 16 / 20 Passed |