Add hosted evaluations section to eval docs

[xeophon]

Summary

• add a short Hosted Evaluations section near the top of docs/evaluation.md
• link the new section from the table of contents
• document the basic prime eval run --hosted flow, including publishing first, --follow, TOML config usage, and the official hosted eval guide

Testing

• Not run (not requested)

---

Note

Low Risk
Low risk documentation-only change; no code paths or CLI behavior are modified.

Overview
Adds a new Hosted Evaluations section to docs/evaluation.md, linked from the table of contents, describing how to run prime eval run --hosted against Hub-published environments (including prime env push, --follow, and TOML config usage) and pointing to the official hosted-evals guide for hosted-only flags.

^{Written by Cursor Bugbot for commit aee2394. This will update automatically on new commits. Configure here.}

[cursor]

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

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

[cursor]

Missing skill update for hosted evaluations workflow

Low Severity

The PR adds a new Hosted Evaluations section to docs/evaluation.md documenting the --hosted flag, --follow, and other hosted-only flags, but skills/evaluate-environments/SKILL.md was not updated. The skill file only mentions "hosted eval workflows" in passing (line 63) and the example command there doesn't even use --hosted. The project rule requires that changes to docs/evaluation.md that affect user-facing workflows are reflected in the corresponding skill file.

 

^{Triggered by project rule: BugBot Instructions}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: aee2394a5d

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]

Keep the pushed namespace consistent with the hosted slug

This example fails if someone copies it literally: prime env push my-env publishes under the caller's own namespace, but the next command runs my-team/my-env. Publishing to a team requires an explicit --team <team> on the push, so the doc should either keep the owner the same in both commands or show the team flag; otherwise readers end up targeting a hosted slug they never created.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Explain that hosted TOML configs must use Hub slugs

The new hosted TOML example omits the main difference from the local TOML flow: hosted [[eval]].env_id entries must point at already-published Hub slugs like owner/my-env. Later in this same file, env_id is documented as an environment module name and every config example uses local IDs, so readers following this new snippet will naturally reuse gsm8k/my-env and get hosted runs that cannot resolve the environment.

Useful? React with 👍 / 👎.