fix(cli): require explicit gpu sandbox flag

[elezar]

Summary

Remove the CLI behavior that inferred GPU sandbox mode from the selected sandbox image name. GPU resources are now requested only when the user passes --gpu.

Related Issue

None.

Changes

• Use the explicit --gpu flag as the sole source of SandboxSpec.gpu.
• Remove the image-name GPU detection helper and its dedicated unit tests.
• Update CLI help and README wording so docs no longer describe implicit GPU inference.

Testing

• mise run pre-commit passes
• Focused Rust tests run: cargo test -p openshell-cli sandbox_should_persist
• E2E tests added/updated (not applicable)

Checklist

• Follows Conventional Commits
• Commits are signed off (DCO)
• Architecture docs updated (not applicable)

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[mrunalp]

/ok to test 344395d

[github-actions]

Label test:e2e-gpu applied for 344395d. Open the existing run and click Re-run all jobs to execute with the label set. The run will execute GPU E2E after building the required supervisor image once. The matching required CI gate status on this PR will flip green automatically once the run finishes.