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-loopSKILL.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
- 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。 - 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。 - 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。 - 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。 - 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。
- 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)
- 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)
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
- SKILL.md in the repositoryCheckingskills/research-loop/SKILL.md
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
[](https://verified-skill.com/skills/biociao/dsh-science/research-loop)