Scgpt
ResearchscgptScgpt is an Agent Skill for research published in the companion-inc/feynman repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
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.
QUICK START
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/companion-inc/feynman/blob/HEAD/skills/scgpt/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/scgpt, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
scGPT
Use this skill for single-cell foundation model analysis.
Workflow:
- Record dataset source, organism, modality, preprocessing, gene identifiers, cell labels, and perturbation design.
- Verify model/checkpoint/package availability before running.
- Save AnnData or matrix manifests, preprocessing code, model version, embeddings, predictions, plots, and logs.
- Compare labels or predictions against source metadata, marker genes, perturbation controls, and literature.
- Preserve batch, donor, disease, and assay provenance in every summary table.
Do not present cell-state labels without marker or metadata support.