Add Prime feature flag evaluation client

[d42me]

Adds a Prime CLI feature flag client that evaluates defaults via the platform feature-flags endpoint with CLI version and team context, falling back to defaults when evaluation is unavailable.

---

Note

Low Risk
Low risk: additive client/utilities plus unit tests, with failures explicitly falling back to provided defaults and no changes to auth or existing request flows.

Overview
Introduces feature-flag evaluation support in the Prime CLI via a new FeatureFlagsClient plus convenience helpers evaluate_feature_flags and is_feature_enabled, which POST defaults along with cli_version and optional team_id to /feature-flags/evaluate and merge missing keys back to defaults.

Exports these utilities from prime_cli.__init__ and adds unit tests covering payload context, empty-flag short-circuiting, API-error fallback behavior, and strict boolean handling for is_feature_enabled.

^{Reviewed by Cursor Bugbot for commit b27f8f3. Bugbot is set up for automated code reviews on this repo. 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.}

^{Reviewed by Cursor Bugbot for commit b27f8f3. Configure here.}

[cursor]

Config without client desyncs

Medium Severity

When FeatureFlagsClient or evaluate_feature_flags get a config but no client, a new APIClient uses its own Config for auth and base URL while team_id in the evaluate payload comes from the passed config, so evaluation context can disagree with the request identity.

Additional Locations (1)

• packages/prime/src/prime_cli/feature_flags.py#L43-L53

 

^{Reviewed by Cursor Bugbot for commit b27f8f3. Configure here.}

[willccbb]

@d42me should we fix or OK?