Content
47%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a comprehensive and well-structured skill with excellent workflow clarity, featuring explicit blocking gates, validation checkpoints, and error recovery. However, it suffers significantly from verbosity — the main SKILL.md tries to be both an overview and a complete reference, inlining large tables and detailed policies that should live in reference files. The actionability is moderate: while the workflow and file structure are concrete, the most critical operational details (prompt assembly, analysis framework) are delegated to reference files without inline examples.
Suggestions
Move the large inline tables (presets, Style × Layout matrix, auto-selection signals) to reference files and keep only 2-3 representative examples in the main SKILL.md to dramatically reduce token count.
Add a concrete, complete example of a prompt file (prompts/01-cover-example.md) inline to make the most critical output artifact actionable without requiring reference file lookup.
Consolidate the Image Generation Tools backend resolution section — the 4-level priority cascade with sub-bullets and multiple ⛔ warnings could be a concise decision table or moved to a reference file with only the key rule stated inline.
Trim the Batch Generation Policy and Reference Images sections to essential rules only — Claude can infer standard batch/retry semantics without the current level of specification.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~500+ lines. It over-explains concepts like batch generation policy, image backend resolution (a 4-level priority cascade with sub-bullets), confirmation policy, and reference image handling in exhaustive detail. Many sections (e.g., the full Style × Layout compatibility matrix, the complete presets tables, auto-selection signal tables) could be in reference files rather than inline. Claude doesn't need this level of hand-holding for concepts like file backup naming conventions or slug generation rules. | 1 / 3 |
Actionability | The skill provides concrete file paths, CLI flags, specific workflow steps, and YAML frontmatter examples, which is good. However, it lacks executable code examples — there are no actual prompt file content examples, no complete generation command examples, and the workflow steps reference external files (e.g., `references/workflows/analysis-framework.md`, `references/workflows/prompt-assembly.md`) for the actual actionable content. The prompt assembly — arguably the most critical part — is entirely delegated to a reference file. | 2 / 3 |
Workflow Clarity | The 4-step workflow is clearly sequenced with explicit blocking gates (Step 0 is marked ⛔ BLOCKING, Step 2 is marked ⚠️ REQUIRED), validation checkpoints (verify prompt files exist before generation, image-1 anchor chain before batching), error recovery (retry failed items once), and a backup rule applied throughout. The confirmation policy with explicit skip conditions and the hard gate between Step 2 and Step 3 demonstrate strong workflow discipline. | 3 / 3 |
Progressive Disclosure | The skill has a well-organized references table at the bottom with clear one-level-deep pointers to 15+ reference files. However, the main SKILL.md itself is a monolithic wall of content — the full presets tables (30+ presets), the 12×8 compatibility matrix, the auto-selection signal table, and the detailed backend resolution logic are all inline when they should be in reference files. The balance between overview and detail is poor, with too much detail kept in the main file. | 2 / 3 |
Total | 8 / 12 Passed |