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.
A dense, domain-rich reference skill: nearly every table carries non-obvious thresholds and signal logic, and the composite scoring framework plus output template make it genuinely usable for producing an on-chain report. Weaknesses are the missing data-acquisition layer (APIs named but never specified), an implicit rather than explicit workflow, and a monolithic single-file structure with time-sensitive historical tables inlined. It would score higher with a numbered analysis procedure and the historical/metric reference split into a references/ file.
Suggestions
Add a short numbered workflow section (fetch metrics -> percentile them vs history -> apply composite weights -> emit the report) so the sequence is explicit rather than implied by section order.
Add minimal concrete data-acquisition guidance — example Glassnode/CryptoQuant API endpoints or the exact metric names to request — since the current text only names 'Glassnode / Nansen' in a note.
Move the historical BTC active-address table and the metric-formula reference into references/ files (and mark the dated thresholds as adjustable), keeping SKILL.md as a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense, well-structured reference data Claude does not know precisely (whale tiers, MVRV zones, SOPR/NVT thresholds, stablecoin signals) with almost no padding or explanation of known concepts. Not 5 because time-sensitive dated tables ('2024 bull market | $40k-$100k', '2021 bull market') are inlined rather than quarantined, and a few asides ('Similar to a PE ratio in traditional finance', 'Simply put: ...') could be trimmed; not 3 because the bulk of every section earns its tokens. | 4 / 5 |
Actionability | Gives concrete thresholds (single transfer > 500 BTC, USDT > $50M), formulas (MVRV = market cap / realized cap), a weighted 1-5 composite scoring scheme, and a full output-report template. Not 5 because the data-acquisition step is only gestured at ('requires extra APIs (Glassnode / Nansen)') with no endpoint, command, or field examples, so the workflow is not fully executable end-to-end. | 4 / 5 |
Workflow Clarity | An analysis sequence exists (snapshot metrics -> whale activity -> composite score -> conclusion) and the output format enforces structure, but the steps are implicit across sections rather than a defined procedure, with no validation or sanity-check checkpoints (e.g. cross-checking conflicting signals). Not 4 because there is no explicit ordered workflow the model can follow step by step; not 2 because the composite-score section and output template do impose a clear sequence and the Notes section flags error sources. | 3 / 5 |
Progressive Disclosure | A single ~250-line file with clean section headers but no bundle files at all (references/, scripts/, assets/ absent); inlinable reference material — the historical BTC active-address table, the metric-formula reference, and the whale-tier table — stays in the main file. Not 4/5 because the skill is well over 50 lines and would benefit from splitting detail into references; not 2 because what is inline is well-sectioned and navigable, not a wall of text. | 3 / 5 |
Total | 14 / 20 Passed |