Skip to content
Automation
Skill

/ds-full-pipeline

Full DeepScientist research pipeline: scout → baseline → idea → experiment → analysis → optimize → write → review → finalize. End-to-end autonomous research lifecycle.

From plugin
dr-claw
1.2k122 skills

Claude Code

$ npx -y skills add OpenLAIR/dr-claw --skill ds-full-pipeline --agent claude-code

SKILL.md

ds-full-pipeline.SKILL.md
name: ds-full-pipeline
description: "Full DeepScientist research pipeline: scout → baseline → idea → experiment → analysis → optimize → write → review → finalize. End-to-end autonomous research lifecycle."
license: MIT
metadata:
  author: ResearAI/DeepScientist
  version: "1.0.0"

DeepScientist Full Pipeline

End-to-end autonomous research workflow for: **$ARGUMENTS**

Overview

This skill chains all DeepScientist research stages into a single pipeline:

/ds-scout → /ds-baseline → /ds-idea → /ds-experiment → /ds-analysis-campaign → /ds-optimize → /ds-write → /ds-review → /ds-finalize

Pipeline

Stage 1: Scout

Frame the research problem, survey literature, identify datasets/metrics, discover existing baselines.

/ds-scout "$ARGUMENTS"

**Output:** Problem framing, literature map, baseline shortlist, evaluation contract.

**🚦 Gate 1:** Present the research landscape to the user. Wait for confirmation before proceeding.

Stage 2: Baseline

Reproduce or import the most relevant baseline from Stage 1's shortlist.

/ds-baseline

**Output:** Working baseline with verified metrics, comparability contract.

Stage 3: Idea

Generate concrete research hypotheses based on the literature gaps and baseline analysis.

/ds-idea

**Output:** Ranked candidate ideas with selection rationale.

**🚦 Gate 2:** Present top ideas to the user. Wait for confirmation of which idea to pursue.

Stage 4: Experiment

Implement and run the main experiment for the selected idea.

/ds-experiment

**Output:** Experiment code, results, evidence artifacts.

Stage 5: Analysis Campaign

Run follow-up experiments: ablations, robustness checks, error analysis.

/ds-analysis-campaign

**Output:** Ablation results, robustness data, writing-facing evidence slices.

Stage 6: Optimize (Optional)

If results are promising but not yet strong enough, run algorithm-first iterative improvement.

/ds-optimize

Skip this stage if main experiment results already meet the success criteria.

Stage 7: Write

Draft the paper from accepted evidence.

/ds-write

**Output:** LaTeX paper draft with figures and references.

Stage 8: Review

Run an independent skeptical audit of the draft.

/ds-review

**Output:** Review report with severity-graded feedback.

If review identifies critical issues → fix and re-review (max 2 rounds).

Stage 9: Finalize

Consolidate final claims, limitations, and recommendations.

/ds-finalize

**Output:** Final paper, summary state, resume packet.

Key Rules

  • **Gate checkpoints after Scout and Idea stages.** Do not proceed without user confirmation on research direction and idea selection.
  • **Stages 4-9 can run autonomously** once the user confirms the idea.
  • **Evidence-first writing.** Every claim in the paper must trace to an experiment artifact.
  • **Fail gracefully.** If any stage fails, report clearly and suggest alternatives rather than forcing forward.
  • **Git as memory.** Commit after each stage so progress is durable.

Typical Timeline

| Stage | Duration | Autonomous? | |-------|----------|-------------| | 1. Scout | 20-40 min | Wait for Gate 1 | | 2. Baseline | 15-60 min | Yes | | 3. Idea | 15-30 min | Wait for Gate 2 | | 4. Experiment | 30 min - hours | Yes | | 5. Analysis | 30-60 min | Yes | | 6. Optimize | 0-60 min | Yes (optional) | | 7. Write | 30-60 min | Yes | | 8. Review | 15-30 min | Yes | | 9. Finalize | 10-20 min | Yes |

Read more
@openlair
Ships withdr-claw

A Super AI Lab with massive AI Doctors as Assistants. Best IDE for Research via AI Power.

1.2k122 skills
Get the whole plugin

Stars on GitHub

Momentum

stars in total
1.2k
Star history of OpenLAIR/dr-claw on GitHub, a chart by star-history.com

Other skills on dr-claw.

dr-claw

dr-claw

Dr. Claw skill for OpenClaw project discovery, idea intake, waiting-session triage, structured session control, event-driven…

@openlair@openlairInstall Skill
academic-researcher

academic-researcher

Academic research assistant for literature reviews, paper analysis, and scholarly writing. Use when: reviewing academic papers,…

@openlair@openlairInstall Skill
aris-ablation-planner

aris-ablation-planner

Use when main results pass result-to-claim (claim_supported=yes or partial) and ablation studies are needed for paper submission.…

@openlair@openlairInstall Skill
aris-arxiv

aris-arxiv

Search, download, and summarize academic papers from arXiv. Use when user says "search arxiv", "download paper", "fetch arxiv",…

@openlair@openlairInstall Skill
aris-auto-review-loop

aris-auto-review-loop

Autonomous multi-round research review loop. Repeatedly reviews via Codex MCP, implements fixes, and re-reviews until positive…

@openlair@openlairInstall Skill

More skills from other plugins.

agent-builder

agent-builder

Load immediately after an Agent intent, before the first agent_builder_* call. Governs how you create, configure, test, and…

n8n12d209k
@n8n-io@n8n-ioInstall Skill
browser-use

browser-use

Direct browser control via CDP for web interaction: automation, scraping, testing, screenshots, and site/app work. Also covers…

@browser-use@browser-useInstall Skill
agent-browser

agent-browser

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling…

@vercel-labs-2@vercel-labs-2Install Skill
add-anydoc

add-anydoc

Add local office-document-to-Markdown conversion to NanoClaw agent containers with the pinned Firecrawl AnyDoc CLI. Use when…

@nanocoai@nanocoaiInstall Skill

Questions about ds-full-pipeline.

Which coding agents can use ds-full-pipeline?
ds-full-pipeline installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.