VMamba Env Doctor

v1.0.0

Guide for setting up a stable VMamba environment and resolving common CUDA, PyTorch, MMCV, and compilation errors. Use when users encounter issues like "unsu...

0· 46·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The name/description match the content: SKILL.md is a step-by-step environment and build guide for VMamba, PyTorch, MMCV and compiling selective_scan. All requested actions (conda/pip/mim installs, setting CC/CXX, building a local kernel) are expected for that purpose.
Instruction Scope
Instructions are narrowly scoped to creating a conda env, installing specific package versions, fixing MKL/NumPy/OpenCV issues, and compiling a local CUDA kernel. The skill does not instruct reading unrelated files, exporting secrets, or contacting unexpected endpoints; network activity is limited to conda/pip/mim package repositories (expected).
Install Mechanism
This is instruction-only (no install spec). The guide tells users to run conda/pip/mim commands which are standard package managers. Nothing is downloaded from obscure personal URLs by the skill itself.
Credentials
No environment variables, credentials, or config paths are required by the skill. The guidance suggests setting CC/CXX for local compilation (expected and limited in scope).
Persistence & Privilege
always:false and no persistent modifications to agent/system configuration are requested by the skill. It does not ask to enable itself or modify other skills.
Assessment
This skill is a coherent, copyable setup guide for VMamba; it does not request credentials or hide endpoints. Before running commands, review them and run them in an isolated conda environment. Confirm your system GPU drivers support CUDA 11.8, avoid running commands as root unless necessary, and inspect any local code you build (the pip install . step in kernels/selective_scan compiles native code). Be aware that mim/pip/conda will fetch packages from public repositories—ensure your environment/trust policies allow those sources.

Like a lobster shell, security has layers — review code before you run it.

latestvk97f092gd0thfsg3f0nh62sefs84swga

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments