Content
77%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.
Operationally excellent: concrete commands, complete examples, and well-sequenced workflows with explicit validation feedback loops. The weaknesses are a ~100-line block of design-taste content Claude could generate itself and a broken reference layer where the two MANDATORY reference files (html2pptx.md, ooxml.md) and the ooxml/scripts/ path are absent from the bundle.
Suggestions
Move the color palette library and visual design options (~100 lines) into a separate reference file (e.g., design.md) with a brief pointer, or trim to 3-4 example palettes — this content is design taste Claude already possesses.
Add html2pptx.md and ooxml.md to the bundle or fix their paths, since both creation and editing workflows are built around a MANDATORY full read of files that do not currently exist.
Clarify the ooxml/scripts/ location for unpack.py, validate.py, and pack.py — the directory is absent from the bundle, and the `find . -name "unpack.py"` fallback only partially mitigates a broken path in a core workflow step.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The operational sections (workflows, commands, XML structure) are tight, but roughly 100 lines are design-taste content Claude already knows — 18 hex-coded palettes ("Classic Blue: Deep navy (#1C2833), slate gray (#2E4053)..."), "Rotated text headers at 90° or 270°", "Minimal gridlines or none at all" — plus heavy emphasis-word repetition (CRITICAL/IMPORTANT/MANDATORY appears on 18 lines). This is a substantial trimmable chunk rather than anchor 4's 'minor instances', but the absence of concept tutorials keeps it above anchor 2. | 3 / 5 |
Actionability | Fully executable throughout: "python scripts/rearrange.py template.pptx working.pptx 0,34,34,50,52", "soffice --headless --convert-to pdf template.pptx", "pdftoppm -jpeg -r 150 -f 2 -l 5 template.pdf slide", a complete example "paragraphs" replacement JSON, and even example validation error outputs. Copy-paste ready commands cover the common cases and edge cases (page ranges, thumbnail grid limits), matching anchor 5. | 5 / 5 |
Workflow Clarity | All three workflows are explicitly sequenced with feedback loops for fragile XML manipulation: "Validate immediately after each edit and fix any validation errors before proceeding: python ooxml/scripts/validate.py <dir> --original <file>" and "Repeat until all slides are visually correct". This mirrors the anchor-5 example, and the destructive/batch validation cap does not apply since validation checkpoints are present. | 5 / 5 |
Progressive Disclosure | References are one level deep and clearly signaled (MANDATORY read directives), and all "scripts/*.py|js" references resolve to real bundle files — but both mandated reference files "html2pptx.md" and "ooxml.md" are missing from the bundle, and the "ooxml/scripts/unpack.py" path does not exist (only a `find` fallback is offered). Combined with ~120 lines of inlined inventory/replacement JSON schema detail, this exceeds anchor 4's 'minor organization gaps' but stops short of anchor 2's minimal-structure profile. | 3 / 5 |
Total | 16 / 20 Passed |