biociao/dsh-science/research-loop

Claude Code and agent skill by biociao/dsh-science

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

执行 Claude Science 式研究循环(ReAct:提问→假设→实验→观察→分析→结论→迭代)。当开始或继续一个科研任务、需要把研究推进到下一轮迭代时使用。配合 research_* 工具维护 research-manifest.json。

Install

npx vskill@latest install biociao/dsh-science/research-loop

SKILL.md

The same SKILL.md was published earlier as blake-john/dsh-scientist/research-loop.

研究循环(Research Loop, ReAct)

研究循环是科研项目的核心执行协议。每一轮迭代 = 一个完整的 Reason → Act → Observe 循环,结果落盘到研究清单(research-manifest.json),跨会话持续。

循环步骤

  1. Question — 明确本轮要回答的问题。用 research_state 读取当前状态;问题不清晰时先用文献调研(literature-connector)收窄。
  2. Hypothesis — 提出可证伪假设,用 research_hypothesis 登记(H1/H2/…)。好假设:可检验、有明确预期方向。
  3. Experiment — 用 research_experiment 登记实验(E01/…),写清 design.md:目的、步骤、预期结果。先写设计再写代码。
  4. Act — 写代码运行。代码放 experiments/<id>/code/,结果放 experiments/<id>/results/。每一步跑什么命令、参数是什么,写进 log.md。
  5. Observe — 对照 design.md 的预期,记录实际观察(数值、图、表)。不要在此步下结论。
  6. Analyze / Conclude — 用 research_findings 记录发现;conclusion 更新假设状态:
  • supported:数据支持假设(注明统计方法与显著性)
  • refuted:数据否定假设(说明哪个预期被违反)
  • inconclusive:无法判定(说明缺什么数据/检验)
  1. Next Question — research_findings 的 nextQuestion 开启下一轮:更新研究问题、迭代数 +1。

阶段推进

用 research_phase 推进:literature → hypothesis → experiment → analysis → manuscript → concluded。结论证据不足时不要跳到 manuscript。

循环纪律

  • 每轮迭代结束必须调用一次 research_findings(即使结论是 inconclusive),保证清单可追溯。
  • 重要的中间产物(数据表、图、模型)立即用 artifact_save 归档(见 artifact-provenance)。
  • 一个会话内推进多轮迭代时,用 todo_write 跟踪每轮步骤;跨会话的长研究目标用 create_goal 建立目标,让目标轮次驱动循环。
  • 关键论断进入论文前必须评审(见 scientific-reviewer)。
  • 遇到阻塞(数据缺失、工具问题)先记录到 log.md,再决定降级方案,不要静默跳过。

示例节奏(一个会话)

research_state → 明确问题
research_hypothesis → 登记假设
research_experiment → 创建 E01
写代码 → 运行 → 记录 log.md
research_findings (finding + conclusion + nextQuestion)
research_phase → 推进阶段

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
45
Repository forks
4
7-day trend score
0

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/biociao/dsh-science/research-loop/badge)](https://verified-skill.com/skills/biociao/dsh-science/research-loop)

Something wrong with this skill or its page? Tell us.