下載技能 ZIP
在 Claude 中上傳
前往 設定 → 功能 → 技能 → 上傳技能
開啟並開始使用
測試它
正在使用「hypothesis-generation」。 Generate hypotheses for why certain cancer patients respond to immunotherapy while others do not.
預期結果:
- Hypothesis 1 (Blue): Tumor mutational burden determines response - patients with high TMB have more neoantigens that immune cells can recognize
- Hypothesis 2 (Green): T-cell inflamed microenvironment - pre-existing T-cell infiltration indicates receptive immune environment
- Hypothesis 3 (Purple): Microbiome composition - specific gut bacteria enhance checkpoint inhibitor efficacy
- Key prediction: If microbiome hypothesis is correct, fecal transplant from responders should transfer responsiveness
- Discriminating experiment: Compare gene expression profiles of responders vs non-responders to identify signature
正在使用「hypothesis-generation」。 Create hypotheses for why certain plants survive drought conditions better than others of the same species.
預期結果:
- Hypothesis 1 (Deep Blue): Deep root system access - superior water retrieval from lower soil layers during drought periods
- Hypothesis 2 (Forest Green): Stomatal regulation efficiency - enhanced leaf-level water conservation through rapid stomatal closure
- Hypothesis 3 (Royal Purple): Osmotic adjustment capacity - cellular solute accumulation maintains turgor under water stress
- Testable prediction: Deep root hypothesis predicts positive correlation between root depth and survival rate under drought
- Experimental design: Controlled drought trial with root imaging to measure depth correlation
安全審計
安全This is a pure documentation skill containing only markdown files, LaTeX templates, and reference guides. No executable code exists. All 273 static findings are false positives: backticks in markdown code blocks were misidentified as shell commands, RGB color definitions were misidentified as cryptographic algorithms, and file references were misidentified as reconnaissance.
風險因素
⚙️ 外部命令 (152)
🌐 網路存取 (7)
品質評分
你能建構什麼
研究提案開發
為論文提案開發嚴謹的假設,配合適當的實驗設計和文獻基礎。
補助提案假設
為資助申請構建具有明確預測的令人信服的可測試假設。
假設驅動分析
在確認性統計檢驗之前,根據探索性數據分析制定假設。
試試這些提示
使用假設生成技能為以下現象建立 3-5 個可測試假設:[描述觀察結果或數據模式]。包括機制解釋、文獻支持證據,以及測試每個假設的實驗設計。
應用假設生成技能來解釋 [現象] 的機制。生成競爭性機制假設,設計區分它們的實驗,並為每個假設提出具體的量化預測。
使用假設生成來開發關於 [研究問題] 的新假設。每個假設都應以 PubMed 文獻證據為基礎,識別當前理解的空白,並設計能推進該領域的實驗。
使用假設生成技能產生全面的假設報告。包括:帶有機制解釋的 3-5 個競爭性假設、琥珀框中的可測試預測、顯示如何區分假設的比較框,以及附錄中詳細的實驗設計。使用 LaTeX 模板格式。
最佳實務
- 從全面的文獻搜索開始,在制定假設之前將其建立在現有證據之上
- 確保每個假設做出明確的、可測試的預測,能夠將其與替代方案區分開來
- 為每個預測包含證偽標準以保持科學嚴謹性
避免
- 產生沒有具體量化預測的模糊假設
- 忽視矛盾證據或替代解釋
- 跳過實驗設計階段 - 假設必須是可測試的