Content
67%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 highly actionable with concrete commands, specific selectors, and well-sequenced workflows that include validation and retry loops. Its main weakness is conciseness — repeated boundary statements and a duplicated URL section — alongside a few placeholder selectors.
Suggestions
Consolidate the '不开独立 session、不 import cookie' boundary into one authoritative statement; the current ~8 repetitions inflate token cost without adding information.
Remove the duplicate 'Feed 详情页 URL 格式' section (lines 75-79) since the same format is already documented in the '获取 feed_id 和 xsec_token' section.
Replace placeholder selectors like '<send-btn-selector>' and the 'camoufox-cli ... open' abbreviation with concrete values so every command is copy-paste ready.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly operational with no padded concept explanations, but the '不开独立 session、不 import cookie' boundary is restated ~8 times and the feed URL format is duplicated across two sections, which is more than minor trimming. | 3 / 5 |
Actionability | Concrete camoufox-cli commands with specific selectors (.content-input, .like-wrapper, .interactions .reply) and eval JS snippets are executable, but a few placeholders ('<send-btn-selector>', 'camoufox-cli ... open') leave minor gaps. | 4 / 5 |
Workflow Clarity | Each interaction type has a clear numbered sequence with snapshot/eval validation checkpoints and retry feedback loops; some confirms rely on visual snapshot judgment rather than explicit pass/fail criteria, a minor validation gap. | 4 / 5 |
Progressive Disclosure | Well-organized into clear sections (login, token extraction, workflows, pitfalls, error table) with content appropriately inline and no nested references, though the single ~200-line file is slightly over-long for a no-bundle skill. | 4 / 5 |
Total | 15 / 20 Passed |