Content
65%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 content is concise, well-organized, and presents a clear end-to-end DJ library workflow. It is held back by a lack of executable code/commands, missing validation checkpoints for batch/destructive metadata operations, and empty bundle directories that limit progressive disclosure.
Suggestions
Add concrete, copy-paste-ready commands for the analyze/tag/export steps (e.g. exact bpm-tools and keyfinder-cli invocations) to lift actionability.
Insert explicit validation checkpoints into the workflow (e.g. 'Verify metadata was written before curating'; 'Confirm crate membership before export') to satisfy the batch/destructive feedback-loop requirement.
Populate references/ with real per-tool reference files (or a detailed USAGE.md) and link to them from the body so detail is offloaded one level deep instead of inlined.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and well-sectioned with no over-explanation of concepts Claude already knows; each section earns its place (purpose, invocation examples, ordered workflow, safety, references). | 5 / 5 |
Actionability | It names concrete tools (bpm-tools, keyfinder-cli, essentia) and gives invocation examples, but provides no executable commands, code, or exact CLI invocations — only high-level guidance like 'Run BPM and key analysis (e.g. via bpm-tools...)'. | 3 / 5 |
Workflow Clarity | A clear 7-step sequence (Ingest → Analyze → Tag → Curate → Playlist export → Journal → Log) is present, but the Safety section lists precautions without validation checkpoints in the workflow; because steps involve batch/destructive metadata writes, the rubric caps workflow clarity at 3. | 3 / 5 |
Progressive Disclosure | Structure is reasonable with clear sections and external reference URLs, but the references/ and scripts/ bundles contain only .gitkeep — there are no real reference files to offload detail to, so content that could be split out (e.g. per-tool usage) is inline and navigation is shallow. | 3 / 5 |
Total | 14 / 20 Passed |