Content
88%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 highly actionable, well-sequenced skill body: exact commands for every mode, an explicit output contract, and error-recovery loops for login-gated platforms. The only notable improvements are trimming time-sensitive literals and moving configuration/troubleshooting detail into reference files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational — no explanations of concepts Claude already knows — but contains minor trimmable/time-sensitive material: the version literal 'v4.0.0-cn' in the badge example, 'Since 2025', and the detailed macOS Catalina browser-failure parenthetical. Not 5 because those tokens don't all earn their place; not 3 because there is no unnecessary explanation, only slight over-detail. | 4 / 5 |
Actionability | Every section provides copy-paste-ready commands with concrete flags and variants — '--emit compact', '--hot', 'login xiaohongshu --cookie ...', '--diagnose --emit json', 'setup' — plus exact env-var configuration. The common cases (topic, hot board, global, diagnostics) are all covered by runnable examples. | 5 / 5 |
Workflow Clarity | The mode-selection table gives an explicit entry point, sections 1–7 sequence mode → run → interpret footer → relay failures, and section 6 forms a clear feedback loop (❌ failure → reason → one-time fix via login/--cookie/--diagnose). Section 5 acts as an interpretation checklist. Not 4: validation and error-recovery checkpoints are explicit rather than 'mostly present'. | 5 / 5 |
Progressive Disclosure | Good structure with the bundle referenced exactly one level deep via '{{SKILL_DIR}}/scripts/last30days.py' (a real file backed by scripts/lib/). Minor gap: everything lives inline in the ~125-line SKILL.md — the Configuration listing and browser/legacy-OS troubleshooting could be split into reference files. Not 5 because the skill exceeds simple-skill scale and some detail-heavy content stays inline; not 3 because organization and signaling are clear. | 4 / 5 |
Total | 18 / 20 Passed |