Content
60%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 core workflow (title validation, PICOS gathering, section generation via the referenced prompts, final assembly) is concrete and well-validated with appropriate offloading to `references/prompts.md`. However, the body is padded with templated filler sections — "When to Use", "Key Features", "Implementation Details" — that add tokens without actionable content and include one broken cross-reference, dragging down conciseness and slightly muddying navigation.
Suggestions
Delete or drastically trim the filler sections ("When to Use", "Key Features", "Implementation Details", "Input Validation") — most bullets are generic templated text; keep only the actual Workflow, Dependencies, and Error Handling content.
Fix the broken pointer "See `## Workflow` above for related details" (Workflow appears below that section) or remove the line entirely.
Make the final step verifiable: replace "Ensure all headings match the PROSPERO requirements" with the concrete heading list or a pointer to where it is defined, and add a check that the word-count constraints (Rationale 5-150, Objectives 10-200) were met.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Noticeably verbose: several padded sections add little value — the templated "When to Use" bullets ("needs a packaged method instead of ad-hoc freeform output"), the generic "Key Features" list, and an "Implementation Details" section that mostly repeats other sections and contains a broken pointer ("See `## Workflow` above" when Workflow appears below). These dilute an otherwise tight core workflow, matching the anchor for several unnecessary padded sections. | 2 / 5 |
Actionability | The Workflow section gives mostly executable guidance: explicit title validation rules with word limit (< 25 words), PICOS gathering, word-count constraints ("Rationale (5-150 words), Objectives (10-200 words)"), runnable commands, and correct pointers to the real prompts in `references/prompts.md`. Minor gaps remain, e.g., "Ensure all headings match the PROSPERO requirements" without stating what those are or where to find them. | 4 / 5 |
Workflow Clarity | The four-step workflow is clearly sequenced with real validation checkpoints — title validation with an explicit retry loop ("If the title fails validation, explain why and ask for a revised title") and missing-PICOS handling. It falls short of a 5 because there is no post-generation verification step (e.g., checking the enforced word counts) and the final "ensure headings match" instruction lacks a concrete check. | 4 / 5 |
Progressive Disclosure | Good structure: the detailed prompts are correctly split into `references/prompts.md` (verified to contain the three referenced prompts) and referenced one level deep with clear signaling ("Use the prompts in `references/prompts.md`"), and `scripts/utils.py` exists as referenced. Minor gaps: the generic "Reference material available in `references/`" filler bullet and the broken "See `## Workflow` above" self-reference slightly impair navigation, keeping it below the clean anchor-5 structure. | 4 / 5 |
Total | 14 / 20 Passed |