Stats
Actions
Available In
Tags
Plugin
airesearchorchestrator
Initialize and run a gated five-phase AI research project from idea through paper. Supports Survey, Pilot, Experiments, Paper, and Reflection phases with explicit human gates.
What's Inside
Slash Commands14
AI Research Orchestrator Commands
/README
This directory contains user-facing commands for the five-phase research workflow.
airesearchorchestrator:abandon
/abandon
Gracefully abandon or archive a research project with proper state cleanup
airesearchorchestrator:configure
/configure
Configure system parameters including research idea, GPU settings, loop limits, and language preferences
airesearchorchestrator:init-research
/init-research
Initialize a new AI research project with proper directory structure and state management
airesearchorchestrator:insight
/insight
Interactive intent clarification — helps users sharpen their research idea before initialization
Agents9
adviser
/AGENT
Reviewer agent for Pilot and Experiments phases. Reviews experimental design, validates results, judges evidence strength.
coder
/AGENT
Primary agent for Pilot and Experiments phases. Designs experiments, implements code, runs experiments, analyzes results.
critic
/AGENT
Reviewer agent for Survey phase. Audits novelty, feasibility, theory risk, and citation authenticity.
curator
/AGENT
Reviewer agent for Reflection phase. Judges which improvements are reusable, safe, and actionable.
orchestrator
/AGENT
Orchestrator agent that coordinates phases, manages state, evaluates gates, and communicates with the researcher. Used by utility skills (configure, status, reload, audit, insight, gitmem, feishu-notify, external-review).
Skills25
airesearchorchestrator:analyze-results
/analyze-results
Analyze ML experiment results, compute statistics, generate comparison tables and insights. Use when user says "analyze results", "compare", or needs to interpret experimental data.
airesearchorchestrator:audit-system-evaluation
/audit-system-evaluation
Audit the system evaluation report for objectivity, evidence quality, and scoring accuracy. Use after Reflector completes system evaluation.
Audit
/audit
This framework is a reference guide, not a rigid checklist. The auditor adapts depth and focus to the specific deliverable, its phase, and what matters most for quality.
airesearchorchestrator:citation
/citation
Unified citation skill covering discovery, verification, ranking, BibTeX generation, and finalization. Finds citation gaps in LaTeX drafts, retrieves real candidate papers, verifies via academic APIs, ranks with a quality formula, and produces provenance-aware BibTeX. Use when user says "curate citations", "verify references", "find citations", "整理引用", "验证引用", or needs citation management for paper writing.
Configure Project Settings
/configure
Configure project-level settings (in `.autoresearch/config/`) and user-level preferences (in `~/.autoresearch/`). Supports interactive wizard and command-line modes.
Stats
Version1.21.0
LanguagePython
Stars1
MaintenanceExcellent
LicenseMIT
Last Commit
Added
Available In
Safety Signals
Caution
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Tags
Categories
