Content
82%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 well-structured, highly actionable workflow body: concrete commands with explicit arguments, a clear phase sequence with validation checkpoints, and thoughtful failure handling for network and API-limit cases. The main gaps are minor duplication in the setup sections, an implicit Phase 2 output check, and referenced bundle files that are not present to verify.
Suggestions
Deduplicate the setup instructions: the 执行环境 section and Step 0 both direct reading shared config/TEMP_DIR — keep one authoritative pointer and drop the repeated instructions.
Add an explicit Phase 2 validation step mirroring Phase 1's output check (e.g., confirm the printed note path exists and the file is non-empty before reporting completion to the user).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes competence — it never explains what GitHub Trending or Obsidian are, and every section carries operational detail (period parsing table, script flags, output conventions). Minor trimming is possible: the 执行环境 section and Step 0 overlap on reading shared config/TEMP_DIR, and the opening announcement instruction ("先说一声 '开始抓 GitHub 榜单 🐙'") is flavor, placing it just below the lean-every-token-earns-its-place anchor. | 4 / 5 |
Actionability | Every phase gives copy-paste-ready commands with exact arguments and placeholders ("python3 fetch_trending.py --since weekly --output \"{TEMP_DIR}/github_trending.json\""), concrete period-to-flag mappings ("周榜 → --since weekly"), file naming rules ("2026-W28 GitHub周榜.md"), and edge-case guidance (SSL fallback, GITHUB_TOKEN handling). Commands fully cover the common cases. | 5 / 5 |
Workflow Clarity | Phases are clearly sequenced (Phase 1 fetch/score → 1b discovery → 2 note writing) with an explicit validation checkpoint ("确认 {TEMP_DIR}/github_trending.json 是非空 JSON 数组。为空则看 stderr 诊断") and error-recovery handling for API rate limits ("status=partial... 不要把空结果写成'没有新增项目'"). Phase 2's output check is only implicit (script prints the note path to stdout), leaving a minor validation gap at the top anchor. | 4 / 5 |
Progressive Disclosure | Well-organized sections with clearly signaled one-level-deep references to shared materials ([language selection](../_shared/language.md), [Agent 运行约定](../_shared/agent-runtime.md), SKILL.en.md) and the three scripts invoked by command. However, none of the referenced bundle files (fetch_trending.py, discover_projects.py, write_trending_note.py, shared docs) are present alongside SKILL.md, so the split can't be verified against the actual bundle structure — good structure with minor organization gaps. | 4 / 5 |
Total | 17 / 20 Passed |