RigorPilot Research Skills
Auditable deep learning research with reproducibility and scientific rigor
What You Can Do
RigorPilot is a comprehensive skill suite for deep learning research, providing auditable repository analysis, README-first reproducibility workflows, bounded exploration with evidence-based ranking, and publication-ready scientific documentation. Each skill enforces rigor through documented assumptions, comparability tracking, and explicit decision checkpoints. Use these skills to understand ML repository internals, faithfully reproduce published results, systematically explore novel ideas with proper baselines, and generate standardized research artifacts.
Features
Read-only deep inspection of deep learning repositories, mapping model architectures, training/inference entrypoints, configs, insertion points, and flagging suspicious implementation patterns without modification.
Faithful reproduction prioritizing README guidance, selecting minimal trustworthy targets (inference before evaluation before training), and recording all deviations, assumptions, and environment state.
Bounded experimental work on frozen branches or checkpoints. Gates ideas with explicit scoring, runs smoke tests, and ranks candidates by real evidence before committing to full execution.
Automatic setup of datasets, checkpoints, dependencies, and caches required for selected reproduction or exploration targets without silent protocol changes.
Generates standardized outputs including SUMMARY.md, SCIENTIFIC_CHANGELOG.md, COMPARABILITY_REPORT.md, ANNOTATED_README.md, and machine-readable status.json for publication-quality evidence.
Records all executed commands, environment variables, code modifications, human decision points, and reproducibility assumptions with conservative patch governance rules.
Loads contextual principles for research safety, deep learning experiments, comparability boundaries, and patch impact to ensure decisions meet scientific standards and align with repository intent.
Example Output
Example 1: Reproducing Baseline Results
I want to reproduce the baseline inference results from the README. What's the smallest target and what do I need to set up?
Claude reads the README, extracts documented inference commands, bootstraps required datasets and checkpoints, executes the minimal target with full logging, and produces SUMMARY.md, COMMANDS.md, SCIENTIFIC_CHANGELOG.md, and COMPARABILITY_REPORT.md showing exact environment, deviations, and metrics reproduction.
Example 2: Exploring Architectural Variants
I have a frozen current_research branch with baseline results. Can you propose and smoke-test 3 architectural variants and rank them by expected impact?
Claude analyzes the repository, gates candidate ideas against the frozen SOTA reference and evaluation method, ranks each by cost and success likelihood, executes one as a smoke test, collects evidence, and documents all decisions with rollback instructions in explore_outputs/.
What's Included
- ai-research-reproduction skill: End-to-end workflow for README-first trusted reproduction with minimal target selection, auditable execution, and standardized documentation
- ai-research-explore skill: Bounded candidate exploration with idea gating, evidence ranking, and scientific comparability tracking for novelty assessment
- analyze-project skill: Read-only repository analysis for architecture, entrypoints, config relationships, and suspicious patterns
- Research governance policies: Embedded references for research rigor, deep learning experiments, patch safety, and continuous learning guidance
- Standardized artifact templates: Pre-configured markdown and JSON formats for SUMMARY.md, SCIENTIFIC_CHANGELOG.md, COMPARABILITY_REPORT.md, and status.json
- Orchestration helpers: Python utilities and scripts for artifact writing, experiment orchestration, and lesson store integration
Who It's For
- Deep Learning Researchers
- ML Paper Authors
- Research Engineers
- Benchmark and Evaluation Specialists
Best For
- Reproducible research workflows
- Auditable experiment exploration
- Scientific rigor and documentation
- Repository-grounded ML analysis
- Publication-ready evidence generation
You might also like
Design and build animations using a proven decision sequence covering when to animate, which tools to use (CSS, WAAPI, Motion.dev), property selection, easing curves, and accessibility. Review existing animations and audit codebases for quality issues. Understand and name motion effects using a comprehensive vocabulary glossary. Learn from Emil Kowalski's opinionated motion design philosophy built on decades of interface engineering experience.
This skill generates customized countdown timelines tailored to your launch type, complete with pre-assigned stakeholder handoffs and clear approval gates. You get built-in risk matrices that surface potential delays before they happen, plus go-no-go decision frameworks that keep your team aligned. Launches execute on schedule, cross-team confusion disappears, and you ship with confidence.
Tackle specialized engineering problems with expert Claude skills spanning CI/CD automation, infrastructure security, open source governance, browser scripting, and documentation workflows. Get authoritative answers to GitHub Actions, Linux hardening, and project sustainability questions grounded in official docs and best practices. Automate routine tasks like userscript development, README localization, and configuration management across diverse technical domains.
This comprehensive collection includes 79 dedicated agent skills for all major Azure services and operations. You can automate Azure resource deployment and management, query infrastructure and logs with natural language using Kusto Graph and IRQL, analyze security events and detect threats, configure identity and access management through Entra ID, monitor application performance with Application Insights, manage storage and messaging services, execute cloud migration workflows, validate compliance policies, and query Azure resources by metadata or properties.
Elevate your interface design from strategy to pixel-perfect execution. Shape layouts and user flows before coding, critique designs against established heuristics, audit for accessibility and performance gaps, and polish for production. Enhance with purposeful animation, strategic color, harmonious spacing, and delightful moments. Includes comprehensive guidance for websites, dashboards, apps, components, forms, onboarding, and design systems across web and native platforms.
by pbakausAccess Vercel's official collection of agent skills for building and deploying modern web applications. Optimize React component architecture using composition patterns to avoid boolean prop proliferation and build scalable component libraries. Deploy applications to Vercel with flexible workflows supporting git integration, CLI authentication, and multiple fallback methods. Apply React and Next.js performance guidelines from Vercel Engineering to write efficient code and improve Core Web Vitals.
Discover Reddit threads where people are genuinely shopping for what you sell, then draft replies that sound like a helpful community member (not an ad). The skill runs two moves: discovery scans target subreddits for recommendation asks, expressed pain, and competitor mentions, ranking threads by signal, product fit, and timing; drafting writes experience-based replies that respect community norms, disclose your affiliation honestly, and only mention your product when it truly answers the question.
Access 29 production-ready Claude agent skills that wrap RunComfy's AI models for image generation, video creation, audio synthesis, and avatar videos. Skills include music generation (ACE Step with inpaint/outpaint), AI avatars and lip-sync (OmniHuman, Seedance, HappyHorse), image editing (inpainting, outpainting, face-swap, relight), video generation and extension, and upscaling. All skills are CLI-driven via the runcomfy command, with built-in schema validation, error handling, and examples for each model.