Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
Highly actionable with a clear, validated multi-phase workflow and genuine one-level references, but the body is verbose with duplicated script examples and generic guidance that competes with context unnecessarily.
Suggestions
Remove the 'Tools and Scripts' usage blocks that duplicate the commands already shown inline in each Phase, or replace them with one-line pointers to the scripts.
Move the generic 'Best Practices' and 'Common Pitfalls to Avoid' sections into a reference file, since much of that guidance is knowledge Claude already has.
Trim the inline BibTeX entry-type reference and the 'Citation Outputs'/'Summary' recap sections, linking to assets/bibtex_template.bib and references/bibtex_formatting.md instead.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~1000-line body duplicates script usage (shown in each Phase and again in 'Tools and Scripts'), explains concepts Claude already knows (BibTeX entry types, generic Best Practices/Common Pitfalls), and could be substantially tightened. | 2 / 3 |
Actionability | Provides many concrete, copy-paste-ready bash commands with real script names and flags (e.g. 'python scripts/extract_metadata.py --doi 10.1038/s41586-021-03819-2'), fully executable. | 3 / 3 |
Workflow Clarity | Five clearly sequenced phases with explicit validation checkpoints and feedback loops (validate → auto-fix → re-validate), satisfying the destructive/batch validation requirement. | 3 / 3 |
Progressive Disclosure | References are real, one level deep, and well-signaled, but large inline sections (Example Workflows, Best Practices, Common Pitfalls) and duplicated script-usage blocks could be split into the existing references/ files for a leaner overview. | 2 / 3 |
Total | 10 / 12 Passed |