In Silico Perturbation Oracle
v1.0.0Virtual gene knockout simulation using foundation models to predict transcriptional changes
⭐ 0· 27·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name and description (virtual gene knockout prediction using foundation models) match the provided code, configs, and SKILL.md. The files include model adapters for 'geneformer' and 'scgpt', configuration files for model and cell-type mappings, and a script to run simulations — all expected for this purpose.
Instruction Scope
SKILL.md instructs installation of ML/bioinformatics packages and running scripts that perform in‑silico perturbation and downstream analyses. The runtime instructions do not request unrelated files, credentials, or system-wide configuration. One minor mismatch: configs reference ${MODEL_DIR} model paths but no MODEL_DIR env var is declared in metadata — users will need to provide or set model download paths when using real models.
Install Mechanism
There is no automated install spec in registry metadata; SKILL.md recommends pip installs (torch, transformers, geneformer, scgpt, etc.). Using pip to fetch heavy ML/model packages is expected for this use case but pulls code from public package repositories — validate package names and sources before installing and prefer isolated environments (venv/conda).
Credentials
The skill declares no required environment variables, credentials, or config paths. The only implicit requirement is a location for pretrained models (configs use ${MODEL_DIR}), which is reasonable for model-based tools. No unrelated secrets or cloud credentials are requested by the bundle.
Persistence & Privilege
The skill is not always-enabled and does not request elevated privileges or permanent presence. It does not modify other skills' configurations. Autonomous invocation is allowed by platform default but is not elevated here and is appropriate for an invocable tool.
Assessment
This package appears coherent for local in‑silico perturbation simulations, but take these precautions before installing or running it: 1) Install and run in an isolated environment (virtualenv or container) because it pulls heavy ML/bio packages. 2) Verify the provenance of third‑party packages (geneformer, scgpt) on PyPI/GitHub before pip installing. 3) Provide or set MODEL_DIR for pretrained model files if you plan to run real models; the configs reference ${MODEL_DIR} but the skill metadata doesn't declare it. 4) Do not run on sensitive or patient-identifiable data without institutional review — outputs are simulated and require wet‑lab validation. 5) Review network activity at first run if you require assurance that no unexpected external endpoints are contacted (the provided code appears local/simulated, but installing model packages may trigger downloads).Like a lobster shell, security has layers — review code before you run it.
latestvk97efad2v4dj44azq27jv5hmxn842q0j
License
MIT-0
Free to use, modify, and redistribute. No attribution required.