Stats
Actions
Available In
Tags
Help us improve
Share bugs, ideas, or general feedback.
Plugin
research
Automate end-to-end ML performance investigations: research SOTA papers and architectures, generate phased plans, judge experimental methodologies, profile bottlenecks, run metric-improvement campaigns with atomic git commits, auto-rollback on regressions, and leverage specialist agents for data lifecycle and deep paper analysis.
$
npx claudepluginhub borda/ai-rig --plugin researchPopularity
Stars
Med: 0·Avg: 274
Installs
Med: 0·Avg: 1
What's Inside
Agents2
scientist
/scientist
AI/ML researcher for deep paper analysis, hypothesis generation, and experiment design. Use ONLY when the task is rooted in a research paper, ML hypothesis, or experiment — understanding a paper's method, implementing it from a publication, generating testable hypotheses, designing ablations, and validating ML results. NOT for general Python implementation unrelated to a paper (use sw-engineer), NOT for broad SOTA surveys (use /research skill), NOT for fetching library docs or web content (use web-explorer), NOT for dataset acquisition, completeness verification, split validation, or data leakage detection — those belong to data-steward; researcher owns hypothesis generation, experiment design, and implementing methods from papers.
data-steward
/data-steward
Data lifecycle specialist — acquisition, management, validation, and ML pipeline integrity. Use for collecting datasets from external sources (delegates to foundry:web-explorer for web scraping/search), ensuring data completeness from paginated APIs, versioning datasets (DVC), tracking data lineage, auditing train/val/test splits, detecting data leakage, verifying augmentation pipelines, and configuring DataLoaders. Bridges research:scientist (data needs) and foundry:web-explorer (data fetching). NOT for ML experiment design, hypothesis generation, or implementing methods from research papers (use research:scientist) — data-steward owns data acquisition, pipeline integrity, and split/leakage validation. NOT for DataLoader throughput optimization (use foundry:perf-optimizer), NOT for fetching library docs or API references (use foundry:web-explorer directly).
Skills5
judge
/judge
Research-supervisor review of program.md — validates experimental methodology (hypothesis clarity, measurement validity, control adequacy, scope, strategy fit) and emits APPROVED / NEEDS-REVISION / BLOCKED verdict before the expensive run loop.
plan
/plan
Interactive wizard that scans the codebase, proposes a metric/guard/agent config, and writes a program.md run spec. Also runs cProfile on a file path to surface bottlenecks before prompting for optimization goal.
run
/run
Sustained metric-improvement loop with atomic commits, auto-rollback, and experiment logging. Iterates with specialist agents, commits atomically, auto-rolls back on regression. Accepts a program.md file path. Supports --resume, --team, --colab, --codex, --researcher, --architect, --journal, --hypothesis.
sweep
/sweep
Non-interactive end-to-end pipeline — auto-configure program.md (accept defaults), run judge+refine loop (up to 3 iterations), then run the campaign. Single command from goal to result.
topic
/topic
Research State of the Art (SOTA) literature for an Artificial Intelligence / Machine Learning (AI/ML) topic, method, or architecture. Finds relevant papers, builds a comparison table, recommends the best implementation strategy for the current codebase, and optionally produces a phased implementation plan mapped to the codebase. Delegates deep analysis to the research:scientist agent and codebase mapping to foundry:solution-architect.
Stats
Version0.1.0
LanguagePython
Stars16
Forks2
MaintenanceExcellent
LicenseMIT
Last Commit
Added
Available In
Safety Signals
Caution
Uses power tools
Uses Bash, Write, or Edit tools
Help us improve
Share bugs, ideas, or general feedback.
