Content
92%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 a well-structured, highly actionable skill: executable commands, explicit validation bands with retry/degradation feedback loops, and clean one-level-deep references that all resolve. The only weakness is mild verbosity in a couple of explanatory asides that could be tightened without losing meaning.
Suggestions
Tighten the matchTier tier-2 explanation: the parenthetical about structured sources being demoted to tier 1 in code can be cut to one clause without losing the actionable signal.
Compress the statistical justifications (the 0/20 CJK sample and the ~27% missing-URL figure) into a short parenthetical each; the recommendation matters more than the empirics.
The 35-code and 41-code language counts are time-sensitive and could drift if the API changes — consider moving exact counts fully into references/languages.md and keeping only the default-rule inline.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient and assumes Claude's competence, but a few spots could be trimmed — e.g. the long matchTier tier-2 parenthetical ("structured sources ... are demoted to tier 1 in code, so a tier-2 hit is either...") and the 27%/0-of-20 statistical asides, which are justified domain knowledge but wordier than needed. | 4 / 5 |
Actionability | Provides fully executable, copy-paste-ready commands for every intent (e.g. `node "$ACADEMY_SKILL_DIR/scripts/academy-api.mjs" prod searchAll '{"query":"gas fee","lang":"en","limit":3}'`), a runnable stripHtml helper invocation, and concrete markdown output templates covering the common cases. | 5 / 5 |
Workflow Clarity | The How-to-Use section is a clearly numbered sequence (detect language → pick intent + call script → format output → templates) with explicit validation checkpoints (matchTier bands 0–3 and -1) and feedback loops (retry with a clean keyword on weak match, the missing-URL degradation ladder). The destructive/batch cap does not apply to this read-only retrieval skill. | 5 / 5 |
Progressive Disclosure | Clear overview body with a well-signaled one-level-deep reference table ("References (loaded-on-demand)") listing all six files with explicit load conditions; all referenced files exist and none nest further, so navigation is easy and the split is appropriate. | 5 / 5 |
Total | 19 / 20 Passed |