AutoResearchClaw — Autonomous Research Pipeline Skill by [ara.so](https://ara.so) — Daily 2026 Skills collection. AutoResearchClaw is a fully autonomous 23 stage research pipeline that takes a natural language topic and produces a complete academic paper: real arXiv/Semantic Scholar citations, sandboxed experiments, statistical analysis, multi agent peer review, and conference ready LaTeX (NeurIPS/ICML/ICLR). No hallucinated references. No human babysitting. Installation Requirements: Python 3.11+ Configuration Minimum config ( config.arc.yaml ) OpenRouter config (200+ models) ACP (Agent Client Protocol) — no API key needed The agent CLI (e.g. claude ) handles its own authentication. OpenClaw bridge (optional advanced capabilities) Key CLI Commands Gate stages (5, 9, 20) pause for human approval in interactive mode. Pass auto approve to skip all gates. Python API Output Structure After a run, all outputs land in artifacts/rc YYYYMMDD HHMMSS <hash / : Pipeline Stages Reference Phase Stage Name Notes A 1 TOPIC INIT Parse and scope research topic A 2 PROBLEM DECOMPOSE Break into sub problems B 3 SEARCH STRATEGY Build search queries B 4 LITERATURE COLLECT Real API calls to arXiv + Semantic Scholar B 5 LITERATURE SCREEN Gate — approve/reject literature B 6 KNOWLEDGE EXTRACT Extract structured knowledge C 7 SYNTHESIS Synthesize findings C 8 HYPOTHESIS GEN Multi agent debate to form hypotheses D 9 EXPERIMENT DESIGN Gate — approve/reject design D 10 CODE GENERATION Generate experiment code D 11 RESOURCE PLANNING GPU/MPS/CPU auto detection E 12 EXPERIMENT RUN Sandboxed execution E 13 ITERATIVE REFINE Self healing on failure F 14 RESULT ANALYSIS Multi agent analysis F 15 RESEARCH DECISION PROCEED / REFINE / PIVOT G 16 PAPER OUTLINE Structure paper G 17 PAPER DRAFT Write full paper G 18 PEER REVIEW Evidence consistency check G 19 PAPER REVISION Incorporate review feedback H 20 QUALITY GATE Gate — final approval H 21 KNOWLEDGE ARCHIVE Save lessons to KB H 22 EXPORT PUBLISH Emit LaTeX + BibTeX H 23 CITATION VERIFY 4 layer anti hallucination check Common Patterns Pattern: Quick paper on a topic Pattern: Reproducible run with full config Pattern: Use Claude via OpenRouter for best reasoning Pattern: Resume after a failed run Pattern: Programmatic batch research Pattern: OpenClaw one liner (if using OpenClaw agent) OpenClaw auto reads RESEARCHCLAW AGENTS.md , clones, installs, configures, and runs the full pipeline. Compile the LaTeX Output Troubleshooting researchclaw: command not found API key errors Experiment sandbox failures The pipeline self heals at Stage 13 (ITERATIVE REFINE). If it keeps failing: Citation hallucination warnings Stage 23 (CITATION VERIFY) runs a 4 layer check. If references are pruned: This is expected behaviour — fake citations are removed automatically Check verification report.json for details on which citations were rejected and why PIVOT loop running indefinitely Stage 15 (RESEARCH DECISION) may pivot multiple times. To cap iterations: LaTeX compilation errors Out of memory during experiments Key Concepts PIVOT/REFINE Loop : Stage 15 autonomously decides PROCEED, REFINE (tweak params), or PIVOT (new hypothesis direction). All artifacts are versioned. Multi Agent Debate : Stages 8, 14, 18 use structured multi perspective debate — not a single LLM pass. Self Learning : Each run extracts lessons with 30 day time decay. Future runs on similar topics benefit from past mistakes. Sentinel Watchdog : Background monitor detects NaN/Inf in results, checks paper evidence consistency, scores citation relevance, and guards against fabrication throughout the run. 4 Layer Citation Verification : arXiv lookup → CrossRef lookup → DataCite lookup → LLM relevance scoring. A citation must pass all layers to survive.

autoresearchclaw-autonomous-research