Add SDK example environments

[jdchawla29]

Summary

• add source-tree coding and computer-use example environments
• make hud init copy a selected bundled example without remote downloads
• use a real SWE-bench Lite task with a self-contained repository bundle and hidden JUnit grading for the coding example
• demonstrate Chromium, Xvfb, x11vnc, noVNC, and the HUD server under one supervisor in the computer-use example
• document the new initialization flow, exercise both examples in CI, and exclude environments/ from distributions

Validation

• uv run pytest hud/cli/tests/test_init.py -q — 13 passed
• uv run pytest tests -q in environments/coding — 15 passed
• uv run pytest tests -q in environments/cua — 5 passed
• uv run ruff check hud/cli environments
• uv run ruff format --check hud/cli environments

---

Note

Medium Risk
Grading and git-vault logic are security-sensitive for eval integrity, and CI now depends on Docker and template matrix stability; init behavior changes affect all new users.

Overview
This PR ships example environment projects in the repo (environments/coding, with cua tracked via gitignore exceptions) and repoints onboarding from remote GitHub presets to copying those trees—hud init defaults to coding in non-interactive runs, with --template for blank, coding, and cua.

The coding environment is the main addition: an ssh Workspace over a repo whose real Git history is vaulted outside the agent view, grading via worktree diff capture, baseline test path restore, application of hidden test_patch, and fail-to-pass / pass-to-pass scoring from JUnit XML (coding/repo.py, coding/grading.py). It includes SWE-bench Lite flask-4992 / flask-5063 rows in tasks.py, a flask.bundle, Dockerfile.hud, and a broad pytest suite.

CI adds a templates matrix that runs hud init, syncs against the checkout SDK, ruff/pytest when present, builds Dockerfile.hud, and smoke-checks the served control channel. Docs (README, quickstart, cookbooks, CLI reference, capabilities) describe the new init flow and coding grading boundary; the cookbooks/codex-coding example is removed.

^{Reviewed by Cursor Bugbot for commit 616d9cd. Bugbot is set up for automated code reviews on this repo. Configure here.}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ecc9b3cbb4

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 42b34ffa76

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 42b34ff. Configure here.}