Content
63%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 with concrete command mappings, an install check, and sensible edge cases, but it is held back by redundant sections (metrics and use/skip guidance each appear twice), padded conversational templates, and references pointing to non-existent files.
Suggestions
Merge the 'Metrics (verified)' section into the Supported commands tables and fold 'Recommendation' into 'Edge cases' so percentages and use/skip guidance appear only once.
Remove or actually bundle the referenced paths 'docs/resource-evaluations/rtk-evaluation.md' and 'examples/claude-md/rtk-optimized.md' — they do not exist in the skill bundle.
Replace the 'Usage pattern' conversational markdown template with one or two concrete executable examples showing the actual rtk invocation and condensed output.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The command tables are dense and useful, but reduction percentages are stated twice ('92.3% reduction' in Supported commands and again in 'Metrics (verified)'), use/skip guidance is duplicated across 'Edge cases' and 'Recommendation', and 'Limitations' carries time-sensitive stats ('446 stars on GitHub... 30 releases in 23 days') outside any deprecated section. | 3 / 5 |
Actionability | Concrete copy-paste commands throughout: mappings like 'git log → rtk git log', install verification ('rtk --version'), fallback installs ('brew install rtk-ai/tap/rtk', 'cargo install rtk'), and 'rtk gain'. Minor gap: the 'Usage pattern' section is a conversational template ('1. Acknowledge request') rather than executable invocation examples. | 4 / 5 |
Workflow Clarity | 'How it works' gives a clear 4-step sequence (Detect → Suggest → Execute when user confirms → Track) with an explicit install-validation checkpoint and fallback commands, plus edge-case handling. Missing error-recovery if the RTK output is wrong or the command fails, so it sits at anchor 4 rather than 5. | 4 / 5 |
Progressive Disclosure | Sections are clearly headed, but the References section points to 'docs/resource-evaluations/rtk-evaluation.md' and 'examples/claude-md/rtk-optimized.md', which do not exist in the bundle, and detailed metrics plus the CLAUDE.md configuration template are inlined in a 135-line file instead of being split out. | 3 / 5 |
Total | 14 / 20 Passed |