Chengkun Rao

Finds the exact dataset split, preprocessing step, or evaluation protocol a deep-learning repo needs but didn't document—by cross-referencing the original paper and repo files.

Rigor Intake helper for README-first deep learning repo reproduction. Use when the task is specifically to scan a repository, read the README and common project files, extract documented commands, classify inference, evaluation, and training candidates, and return the smallest t…

Parses a README, extracts environment and asset requirements, builds a conda env spec, and generates step-by-step setup notes so the repo actually runs without path errors or missing checkpoints.

Runs a deep-learning repo's smoke test or eval command, captures outputs, and writes them into a normalized repro_outputs/ folder with patch notes and environment metadata.

Reads a GitHub repository's README, picks the smallest reproducible target (inference or eval), installs deps, runs it, and logs every step so you can trust the result.

Takes a traceback, CUDA error, shape mismatch, or NaN loss and walks through the root cause with conservative reasoning before suggesting a fix — no wild patches.

Rigor Train skill for deep learning research repositories. Use when a documented or selected training command should be run conservatively for startup verification, short-run verification, full kickoff, or resume, with command, config, seed, log, checkpoint, status, and metric e…

Maps a machine-learning repository's model architecture, training flow, and config insertion points—without running code. Flags unusual patterns or missing safety checks.

Explores a research repo's architecture, identifies reusable modules, and generates isolated branch changes to swap backbones, add adapters, or integrate LoRA without breaking the main codebase.

Executes bounded guess-and-check experiments—small-subset validation, batch sweeps, short hyperparameter searches—on your deep-learning repo without consuming full compute budgets.

Takes a task family, dataset, benchmark, and state-of-the-art references, then generates candidate research angles with a rationale for each — filtered for novelty and feasibility.

Rigor Analyze / Rigor Audit read-only skill for deep learning research repositories. Use when the user wants to read and understand a repository, inspect model structure and training or inference entrypoints, review configs and insertion points, or flag suspicious implementation patterns without modifying code or running heavy jobs. Do not use for active command execution, broad refactoring, speculative code adaptation, or automatic bug fixing.

Rigor Improve implementation leaf skill for auditable candidate implementation in deep learning research repositories. Use when the researcher explicitly authorizes exploratory work on an isolated branch or worktree to transplant modules, adapt a backbone, add LoRA or adapter layers, replace a head, or stitch together meaningful low-risk migration ideas with rollback-aware records in `explore_outputs/`. Do not use for end-to-end exploration orchestration on top of `current_research`, trusted baseline reproduction, conservative debugging, environment setup, verified contribution claims, or default repository analysis.

Rigor Explore compatible skill slug for meaningful and potentially novel deep learning research candidates. Use when the researcher has chosen the task family, dataset, benchmark, evaluation method, provided SOTA references, and wants candidate-only exploration on top of `current_research` with auditable repo understanding, idea gating, fair comparison, and governed experiments written to `explore_outputs/`. Do not use for README-first trusted reproduction, open-ended direction finding, narrow code-only or run-only exploration, passive repo analysis, verified novelty claims, or implicit experimentation.

RigorPilot reproduce-mode orchestrator for README-first deep learning repository reproduction. Use when the user wants an end-to-end, minimal-trustworthy flow that reads the repository first, selects the smallest documented inference or evaluation target, coordinates intake, setup, trusted execution, optional trusted training, optional repository analysis, and optional paper-gap resolution, enforces conservative patch rules, records evidence assumptions deviations and human decision points, and writes the standardized `repro_outputs/` bundle. Do not use for paper summary, generic environment setup, isolated repo scanning, standalone command execution, silent protocol changes, score chasing, or broad research assistance outside repository-grounded reproduction.

Rigor Paper Context helper for README-first deep learning repo reproduction. Use only when the README and repository files leave a narrow reproduction-critical gap and the task is to resolve a specific paper detail such as dataset split, preprocessing, evaluation protocol, checkpoint mapping, or runtime assumption from primary paper sources while recording conflicts. Do not use for general paper summary, repo scanning, environment setup, command execution, title-only paper lookup, or replacing README guidance by default.

Runs a machine-learning training command from a research repo with checkpoints, logs, and metric snapshots saved to a standard output folder. Handles startup checks, short runs, full kickoff, and resume.

Parses a traceback, CUDA error, or NaN loss and pinpoints the root cause without guessing. Returns only the minimal fix, keeping research logic intact and debug changes clearly marked.

Scans a README and project structure, extracts training/eval/inference commands, and returns an ordered checklist of the smallest steps to reproduce the result—without running anything yet.

Rigor Setup skill for README-first deep learning repo reproduction. Use when the task is specifically to prepare a conservative conda-first environment, checkpoint and dataset path assumptions, cache location hints, and setup notes before any run on a README-documented repository. Do not use for repo scanning, full orchestration, paper interpretation, final run reporting, or generic environment setup that is not tied to a specific reproduction target.

Runs the repo's documented inference or evaluation command, captures outputs, and logs what changed—so you know whether the claims in the README actually hold up.

Executes bounded exploratory runs—small-subset validation, parameter sweeps, transfer-learning trials—on your research repo and logs results with honest caveats about what the data does and doesn't prove.

Main orchestrator for README-first AI repo reproduction. Use when the user wants an end-to-end, minimal-trustworthy reproduction flow that reads the repository first, selects the smallest documented inference or evaluation target, coordinates intake, setup, trusted execution, optional trusted training, optional repository analysis, and optional paper-gap resolution, enforces conservative patch rules, records evidence assumptions deviations and human decision points, and writes the standardized `repro_outputs/` bundle. Do not use for paper summary, generic environment setup, isolated repo scanning, standalone command execution, silent protocol changes, or broad research assistance outside repository-grounded reproduction.