Feat: Add secrets subcommands

[d42me]

Note

Medium Risk
Adds new CLI flows for creating/updating/deleting secrets and variables and changes API client handling of empty/204 responses, which could affect multiple commands if response parsing expectations differ.

Overview
Adds first-class secret and variable management to the CLI: new top-level prime secret commands for global (personal/team) secrets, plus prime env secret and prime env var subcommands with interactive selection, validation, and table/json output.

Refactors shared CLI utilities by centralizing ANSI stripping in utils.formatters, introducing reusable prompt helpers/validation in utils.prompt, and updating the API client to gracefully handle 204/empty responses (sync + async). Also expands env with a new eval runner (run_eval) that validates models against Prime Inference, injects tracking headers, and can auto-push results after successful runs, and tightens the RL W&B key check.

^{Written by Cursor Bugbot for commit 2e3367f. 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.}