Scvi Tools

What it does

When your agent uses it

• You need probabilistic batch correction (scVI)
• Transfer learning
• Uncertainty-aware differential expression
• Multimodal integration (totalVI/MultiVI)

Example prompts

• “/scvi-tools”

Requirements

• Python 3

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Batch correction and dataset integration for scRNA-seq using a probabilistic latent space (e.g., scVI).
2. 2Transfer learning / semi-supervised annotation when you have partial labels or want to map new data onto a reference (e.g., scANVI).
3. 3Uncertainty-aware differential expression where effect sizes and posterior uncertainty matter (Bayesian DE).
4. 4Multimodal integration across RNA+protein (CITE-seq) or RNA+ATAC (multiome), including paired/unpaired settings (e.g., totalVI, MultiVI).
5. 5Specialized modalities such as ATAC-seq, spatial transcriptomics deconvolution/mapping, doublet detection, methylation, or RNA velocity.

What it can do on your machine

Read from SKILL.md and the folder at commit 686e09d. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • uv

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md. Its commands use uv, which can reach the network depending on how they are called.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Scvi Tools loads about 1.5k tokens when it runs, and up to ~26k if it reads all its reference files. Until then it costs about 55 tokens; SKILL.md has 419 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~55

When it runs · the whole SKILL.md, loaded when a task matches

~1.5k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~26k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from aipoch/medical-research-skills at commit 686e09d, republished under its MIT licence (© aipoch). 419 words, ~1,475 tokens.

name

scvi-tools

description

Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI).

license

MIT

author

AIPOCH

uv pip install scvi-tools
# Optional GPU extras (package extra name may vary by platform/version)
uv pip install "scvi-tools[cuda]"

import scanpy as sc
import scvi

# 1) Load example data (AnnData)
adata = scvi.data.heart_cell_atlas_subsampled()

# 2) Minimal preprocessing (keep raw counts available)
sc.pp.filter_genes(adata, min_counts=3)
sc.pp.highly_variable_genes(adata, n_top_genes=1200)

# 3) Register AnnData for scVI (raw counts + covariates)
scvi.model.SCVI.setup_anndata(
    adata,
    layer="counts",                 # raw counts layer (not log-normalized)
    batch_key="batch",              # batch column in adata.obs
    categorical_covariate_keys=["donor"],
    continuous_covariate_keys=["percent_mito"],
)

# 4) Train model
model = scvi.model.SCVI(adata)
model.train()

# 5) Extract outputs
adata.obsm["X_scVI"] = model.get_latent_representation()
adata.layers["scvi_normalized"] = model.get_normalized_expression(library_size=1e4)

# 6) Downstream analysis with Scanpy
sc.pp.neighbors(adata, use_rep="X_scVI")
sc.tl.umap(adata)
sc.tl.leiden(adata)

# Optional: uncertainty-aware differential expression
de = model.differential_expression(
    groupby="cell_type",
    group1="TypeA",
    group2="TypeB",
    mode="change",
    delta=0.25,
)
print(de.head())

model.save("./scvi_model", overwrite=True)
model2 = scvi.model.SCVI.load("./scvi_model", adata=adata)

© aipoch, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 9 other files (references) in scientific-skills/Data Analysis/scvi-tools of aipoch/medical-research-skills.

• SKILL.md
• references/differential-expression.md
• references/models-atac-seq.md
• references/models-multimodal.md
• references/models-scrna-seq.md
• references/models-spatial.md
• references/models-specialized.md
• references/theoretical-foundations.md
• references/workflows.md
• scvi-tools_audit_result_v1.json

Open the folder on GitHubat commit 686e09d