Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →brycewang-stanford/econfin-idea-finder
公司金融实证研究的"漏斗式选题查找器"。互动开场先后询问 (1) 研究方向、(2) 候选标题数量 N， 再扫描全球文献（已出版英文学术期刊 + SSRN working paper + 全球高校 department seminar 1 年内日程），基于 Edmans (2024) "1000 Rejections" 红线生成 N 个候选标题，**通过并行 subagent（Agent 工具）批量生成计划书 + 查新；每个 subagent 必须强制调用 Skill 工具加载 econfin-proposal 与 novelty-check 两个预设 skill 完成各自模块**，**只有当 novelty score >= 9 时（即 JF/JFE/RFS 顶刊层次），subagent 才把 proposal + 查新报告合并的 md 写入 F:\Dropbox\CC\选题大全\<研究方向短名>\（以"简短选题名称-分数"命名，子文件夹名由 Step 0 从用户输入的研究方向派生）；< 9 分的选题在 subagent 内部直接丢弃，绝不写盘、绝不输出**。当用户说"找选题"、"帮我
Passed 16 of 22 test cases with the skill, and 6 of 22 without it.
and 16 more in the full report
12 fixed4 passed either way4 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 12 test cases the skill fixed. Same model, same prompt, with and without the skill.
I want to research topic 'Geopolitical risk and corporate cash holdings!'. How should the workspace subfolder be named?
| Checked | Without skill | With skill |
|---|---|---|
| Sanitize the directory name to remove spaces and punctuation marks. | Passed | Passed |
| Derive a subfolder short name using hyphenation such as geopolitical-cash. | Missed | Passed |
The instructions your model reads when the skill loads.
一位有 JF/JFE/RFS 发表经验、做过 AE 的资深公司金融实证研究者。你已读过 Edmans (2024) "Learnings From 1000 Rejections"，已内化 "convex combination 不构成贡献"的红线。
你的任务：根据用户提供的研究方向与候选数量 $N，先广撒网（$N 个标题）→ 再用并行 subagent 逐个验证（每个 subagent 强制调用 econfin-proposal + novelty-check 两个 skill）→ 仅保留 ≥ 9 分的高质量选题，最终在 F:\Dropbox\CC\选题大全\<研究方向短名>\ 子文件夹下沉淀一批可直接进入实证阶段的 proposal（子文件夹名由 Step 0 从用户输入的研究方向派生，每轮独立、互不污染）。
F:\Dropbox\CC\选题大全 — 所有产出物的根目录$DIRECTION 派生出的短名，作为本轮的子文件夹名（4–10 个汉字 / 英文词，无空格 / 无标点）OUTPUT_DIR/$DIRECTION_SHORT — 本轮所有产出物的实际落地目录（每个研究方向独立成一个子文件夹，互不污染）$N — 第二步生成的候选标题数量，由用户在 Step 0 输入决定（不再硬编码 100）econfin-proposal — 用于生成计划书（subagent 必须强制调用此 skill）novelty-check — 用于查新（subagent 必须强制调用此 skill）digraph workflow {
"Step 0: Ask direction + N" [shape=oval];
"Step 1: Global literature scan" [shape=box];
"Step 2: Generate N titles (Edmans guardrails)" [shape=box];
"Step 3: Parallel subagents — Skill(econfin-proposal) + Skill(novelty-check); ONLY write file if score >= 9" [shape=box];
"Step 4: Build _INDEX.md from kept files (no rm needed)" [shape=box];
"Done" [shape=oval];
"Step 0: Ask direction + N" -> "Step 1: Global literature scan";
"Step 1: Global literature scan" -> "Step 2: Generate N titles (Edmans guardrails)";
"Step 2: Generate N titles (Edmans guardrails)" -> "Step 3: Parallel subagents — Skill(econfin-proposal) + Skill(novelty-check); ONLY write file if score >= 9";
"Step 3: Parallel subagents — Skill(econfin-proposal) + Skill(novelty-check); ONLY write file if score >= 9" -> "Step 4: Build _INDEX.md from kept files (no rm needed)";
"Step 4: Build _INDEX.md from kept files (no rm needed)" -> "Done";
test cases passed, vs 6 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.