git:20260925.6f9cd81 to git:20260926.2d8ce6a

6 added, 0 removed. Audit A to A.

---

name: cohesix-ai-host-setup

description: Configure and verify third-party AI tools on a Cohesix macOS Apple Silicon or Linux NVIDIA host. Use for MLX, Modular MAX, vMLX, Hugging Face, PEFT, CUDA, and NeMo Agent Toolkit environment setup or repair, not for claiming Cohesix target acceptance.

license: Apache-2.0

---

# Prepare an AI host

Use this skill when the user asks to set up, repair, or inventory third-party

AI tools alongside Cohesix. Read [Mac setup](references/mac.md) for Apple

Silicon and [Linux setup](references/linux.md) for an NVIDIA host. `MAX` means

Modular MAX; `MLX` means Apple's array/model stack; `vMLX` is a separate desktop

app. Select only the requested tools and an actual supported host profile.

+ The repository's pinned NeMo Agent Toolkit kit is for Linux AArch64; a Mac

+ agent client needs its own qualified MCP/A2A path.

The external model runtime, agent, caches and credentials live on the host,

not in seL4. The Mac may be a controller without local MLX work; a Linux

controller need not have CUDA. Read the selected installation's

[host-tool contract](https://raw.githubusercontent.com/lukeb-aidev/cohesix/main/docs/HOST_TOOLS.md),

[GPU contract](https://raw.githubusercontent.com/lukeb-aidev/cohesix/main/docs/GPU_NODES.md), and

[Python support](https://raw.githubusercontent.com/lukeb-aidev/cohesix/main/docs/PYTHON_SUPPORT.md)

when attaching Cohesix tools. Pin these `main` URLs to the selected source

commit/tag, or read the same documents in that checkout.

The repository's `toolchain/setup_*` scripts establish its own build dependencies;

they are not an AI-runtime installer. Preserve the seL4 profile environment and

the repository's hash-locked `.venv` unless its owner explicitly changes them.

## Decide from the live host

1. Record OS release, architecture, Python interpreters, accelerator and driver

or JetPack identity, free disk, model/cache locations, package managers,

installed app/CLI/package/container versions, and actual command paths.

Compare this with the operator's inventory; treat older entries as leads.

2. Select the workload and its compatible runtime *before* installation:

local Mac MLX, optional MAX, vMLX model/client, Linux CUDA + HF PEFT, or

NeMo Agent Toolkit MCP/A2A client. Consult the selected upstream release's

compatibility matrix. Pin the exact package versions, image digests, model

revisions and adapter/base pairing in the environment record. Do not turn

an unqualified `latest` tag or nightly into a reproducible profile.

+ If the requested OS, architecture, accelerator or tool combination is

+ unsupported, name the exact incompatibility and offer a compatible host,

+ runtime or version before making changes. Do not transplant Mac setup

+ commands onto Linux or Linux service and driver instructions onto macOS.

3. Reuse a working, compatible environment. Put a user-wide `hf` CLI on PATH;

install Python model packages in the selected runtime, not in system Python.

Use a separate NeMo client environment only when dependency compatibility

requires it, and expose its `nat` command on PATH. A GPU container is a

distinct native execution runtime, not an incidental Python venv.

4. Before changing drivers, containers, Python packages, apps, services or

caches, inspect dependents, running jobs and retained artifacts. Upgrade or

remove only the selected surface; preserve a working runtime until its

replacement passes the same smoke. Never put tokens, sudo passwords or

signing identities in a tracked file or terminal log.

## Prove the requested capability

Use the platform reference's checks. A version print or successful import is

installation evidence. A small real Metal or CUDA forward/backward step,

adapter load, live model response, or native NeMo client call proves that

specific capability. MCP discovery, A2A task state and model text do not prove

the native job outcome. For a Cohesix workflow, follow the selected

[Test Plan](https://raw.githubusercontent.com/lukeb-aidev/cohesix/main/docs/TEST_PLAN.md)

and the

[GPU operations skill](https://raw.githubusercontent.com/lukeb-aidev/cohesix/main/skills/cohesix-gpu-operations/SKILL.md);

correlate the original operation with signed evidence. Report setup, native smoke,

protocol-client compatibility and live Cohesix acceptance separately.

Return a concise environment record: host/profile and source revision;

component, version and absolute path or immutable image digest; model and

cache location; command/check and observed result; blocker; and the exact

scope of any claim. Include install/repair commands that another operator can

repeat without local credentials or private paths.