Content
56%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 skill's core workflow is genuinely strong: a precise data contract, executable R-first/Python-fallback commands, input validation with a resubmit loop, and a deterministic output template. However, the body is heavily padded with generic scaffold boilerplate, contains a broken cross-reference and contradictory duplicate dependency lists, and buries the real workflow under redundant policy sections, hurting token efficiency and organization.
Suggestions
Delete or collapse the generic template sections ('When to Use', 'Key Features', 'Implementation Details', 'Output Contract', 'Deterministic Output Rules', 'Completion Checklist') that restate nothing specific to forest plots; keep only the Data Format, Workflow, Dependencies, and When Not to Use sections.
Resolve the two contradictory R dependency lists ('ggplot2, meta, gridExtra' vs 'meta, metafor, grid, stringr') into one accurate list, and fix the broken reference 'See ## Workflow above'.
Remove the environment-specific dated path ('cd "20260316/scientific-skills/..."') and point the 'Quick Validation' section at the actual task scripts (forest_binary.R / forest_binary.py) instead of the extract_criteria.py scaffold.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Roughly half the body is generic template filler ('Use this skill when the request matches its documented task boundary', 'Prefer this skill for repeatable, checklist-driven execution'), with a broken self-reference ('See ## Workflow above' when Workflow is below), a hardcoded dated path ('cd "20260316/scientific-skills/..."'), and two contradictory R dependency lists ('ggplot2, meta, gridExtra' vs 'meta, metafor, grid, stringr') — noticeably verbose with several padded sections (anchor 2), though the core Data Format/Workflow sections are efficient, keeping it above anchor 1. | 2 / 5 |
Actionability | Provides copy-paste-ready commands with per-parameter documentation ('Rscript scripts/forest_binary.R "<csv_path>" "<outcome_name>" "<output_dir>"' plus a flagged Python fallback), an exact column table with examples, and a precise output template — mostly executable with minor gaps: no end-to-end example with real values, and the 'validation shortcut' points at the scaffolding script extract_criteria.py rather than the actual task scripts. | 4 / 5 |
Workflow Clarity | A clear 3-step sequence with explicit validation in Step 1 (required columns, 'at least 2 studies, non-negative integer values', with a correct-and-resubmit loop) and an error-recovery fallback in Step 2 ('If R script execution fails, automatically fall back to Python script'). Not anchor 5 because there is no guidance for when both scripts fail, and the scattered duplicate validation sections muddy the primary flow. | 4 / 5 |
Progressive Disclosure | All referenced scripts exist one level deep in scripts/ with no nested references, but everything is inlined in a single ~155-line file whose generic policy sections ('Output Contract', 'Deterministic Output Rules', 'Completion Checklist', duplicated validation sections) are redundant and could be consolidated or dropped; structure exists but organization is inconsistent (anchor 3). | 3 / 5 |
Total | 13 / 20 Passed |