Content
71%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.
The body is well-organized, actionable, and workflow-clear with strong viewport-fitting rules. Its main weakness is progressive disclosure: several referenced supporting files are cited but missing from the actual bundle.
Suggestions
Ship the referenced bundle files (STYLE_PRESETS.md, viewport-base.css, html-template.md, animation-patterns.md) or inline the essential parts so referenced paths resolve.
Trim the motivational 'avoid AI slop' prose to the concrete rules; the design guidance is already conveyed by the preset table and CSS invariants.
Add an explicit validate/fix/retry loop for the generation phase (e.g. open the output and confirm each .slide fits 100vh) to strengthen workflow feedback.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Generally efficient and well-structured by phase, with only minor motivational padding (e.g. 'it is critical that you think outside the box!') that could be trimmed without losing guidance. | 4 / 5 |
Actionability | Provides concrete executable guidance — exact CSS rules, a runnable extraction command, density-limit tables, and AskUserQuestion prompts — with minor gaps where the full HTML template is delegated to a referenced file. | 4 / 5 |
Workflow Clarity | Clear Phase 0–5 sequence with explicit verification checkpoints in Mode C ('After ANY modification, verify...') and density enforcement; minor gaps in failure-recovery feedback loops during generation. | 4 / 5 |
Progressive Disclosure | Structure and the 'Supporting Files' table with a 'When to Read' column are well-signaled and one level deep, but most referenced files (STYLE_PRESETS.md, viewport-base.css, html-template.md, animation-patterns.md) do not exist in the bundle — only scripts/extract-pptx.py is present. | 3 / 5 |
Total | 15 / 20 Passed |