Content
73%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-workflowed skill with concrete executable commands and strong validation loops, undermined by significant repetition (script usage appears four times) and inline content that duplicates the reference files. Trimming the redundancy would markedly improve token efficiency.
Suggestions
Consolidate the four repeated generate_slide_image.py / slides_to_pdf.py usage blocks into one canonical 'Script Reference' section; cross-reference it elsewhere instead of re-explaining.
Move the detailed design-principle content that duplicates references/slide_design_principles.md out of SKILL.md, keeping only a concise summary plus the pointer.
Remove motivational padding and truisms Claude already knows (e.g. 'Scientific presentations are a critical medium for communicating research', 'Boring = Forgotten') to tighten conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Noticeably verbose: generate_slide_image.py usage is explained four times (lines 81-143, 200-260, 980-1010, 1091-1103), the visual-review workflow is duplicated (554-610 and 790-820), design advice repeats across four sections, and motivational padding ('Boring = Forgotten', 'Scientific presentations are a critical medium...') states concepts Claude already knows. | 2 / 5 |
Actionability | Fully executable, copy-paste-ready bash commands with complete prompt templates, formatting goals, citation inclusion, and figure-attach examples covering the common cases (title, content+citations, results with attached figures, comparison, methodology). | 5 / 5 |
Workflow Clarity | Multi-step processes are clearly sequenced with explicit validation: the visual-review workflow has a validate→document→fix→re-validate feedback loop with stopping criteria (558-610), automated validate_presentation.py is provided, and checklists exist (660-673). Validation is present, so the destructive/batch cap does not apply. | 5 / 5 |
Progressive Disclosure | Good structure with a clear overview and well-signaled one-level-deep references (all referenced files in references/ and assets/ exist and are listed in dedicated sections), but substantial design guidance is duplicated inline despite slide_design_principles.md existing, and several sections repeat content already in references. | 4 / 5 |
Total | 16 / 20 Passed |