feynman

Create a research watch baseline and optionally schedule follow-up checks when scheduling tools are visible. Use when the user asks to monitor a field, track new papers, watch for updates, or set up alerts on a research area.

Call a configured Feynman model endpoint and interpret its response. Use when a task needs inference from a registered endpoint, remote model API, local model service, or custom connector-backed predictor.

Compare multiple sources on a topic and produce a grounded comparison matrix. Use when the user asks to compare papers, tools, approaches, frameworks, or claims across multiple sources.

Design or screen protein sequences for solubility-aware constraints with SolubleMPNN-style workflows. Use when a task asks for soluble protein design, expression-friendly variants, or solubility risk filtering.

Create or revise Feynman skills. Use when a research workflow needs a reusable on-demand capability, skill metadata, trigger wording, references, scripts, or skill validation.

Recover prior Feynman work from session transcripts. Use the optional /search command only when it is installed and visible; otherwise search local session JSONL files directly.

Write a durable session log capturing completed work, findings, open questions, and next steps. Use when the user asks to log progress, save session notes, write up what was done, or create a research diary entry.

Inspect the active Feynman workbench session, artifacts, execution log, settings, and provenance. Use when the task asks what happened in this session, which files were written, what tools ran, or what remains unverified.

Run scvi-tools single-cell workflows. Use when a task asks for scVI/scANVI setup, latent embeddings, batch correction, differential expression, cell annotation, or reproducible AnnData analysis.

Use scGPT-style single-cell foundation model workflows. Use when a task asks for single-cell embeddings, perturbation prediction, cell annotation, batch transfer, or gene-program analysis.

Provision and manage GPU pods on RunPod for explicitly chosen long-running research experiments. Use when a Feynman replication, benchmark, or dataset-heavy research run needs persistent GPU compute with SSH access.

Run a tough but constructive internal research critique of an AI research artifact. Use when the user asks for a review, critique, feedback on a paper or draft, or wants to identify weaknesses before submission.

Plan a replication of a paper, claim, or benchmark, and execute only after an explicit environment choice. Use when the user asks to replicate results, reproduce an experiment, verify a claim empirically, or build a replication package.

Run Feynman research jobs on SSH, Slurm, or lab hosts. Use when a task needs remote host setup, job submission, log harvest, artifact sync, or GPU/cluster execution outside Modal.

Dispatch Feynman research notebook or experiment jobs to Modal. Use when a task has explicitly chosen Modal for bounded cloud compute, GPU jobs, or reproducible remote execution.

Design protein sequences from fixed backbone structures with ProteinMPNN-style workflows. Use when a task asks for backbone-conditioned sequence design, mutation suggestions, fixed residues, or design filtering.

Answer questions about Feynman's own product behavior, runtime, settings, commands, skills, connectors, and package state. Use when a response would claim what Feynman can do or how it is wired.

Preview Markdown, LaTeX, PDF, or code artifacts when preview commands are visible, or fall back to shell/browser tools. Use when the user wants to review a written artifact, export a report, or view a rendered document.

Read, extract, and cross-check content across scientific PDFs. Use when a task needs methods, figures, tables, citations, accessions, or claims from multiple places in one or more papers.

Turn research findings into a polished paper-style draft with sections, equations, and citations. Use when the user asks to write a paper, draft a report, write up findings, or produce a technical document from collected research.

Shape the scientific story across a manuscript, abstract, figures, and evidence. Use when a task asks for paper structure, figure order, argument flow, missing analyses, or manuscript revision strategy.

Compare a paper's claims against its public codebase. Use when the user asks to audit a paper, check code-claim consistency, verify reproducibility of a specific paper, or find mismatches between a paper and its implementation.

Run or plan OpenFold3-style structure prediction workflows. Use when a task asks for open protein or complex prediction, setup, model comparison, or reproducibility around OpenFold-family outputs.

Run explicitly chosen research benchmark or replication jobs on Modal's serverless infrastructure. Use when a Feynman research workflow needs burst remote GPU compute and the Modal CLI is available.

Find implementable ML training recipes from papers, datasets, docs, and code. Use when the user wants to fine-tune, train, reproduce, or choose a practical ML method, dataset, hyperparameter setup, or benchmark recipe.