keeper-test

Quarterly roster performance reviews are automated by wrapping roster and performance-reviewer evaluations into a single invocation, using the "would we re-hire" criterion to generate per-agent scores and actionable recommendations.

The "would we re-hire" criterion is an evaluation mechanism that scores agents during performance reviews by assessing their ongoing value, triggering Keep, Upgrade, Repurpose, or Fire actions governed by ADRs for auditability.

On-demand performance reviews trigger from specific events like CAO findings, red-team bouncebacks, or user requests, pulling inputs from sessions and meetings to produce consolidated evaluations with documented justifications.

Yes, performance reviews generate actionable ADRs for Keep, Upgrade, Repurpose, and Fire actions, writing documented justifications and logging outputs to the decisions and vision folders for full traceability.

Roster performance evaluations require input data from vision, meetings, and sessions sources, combining this information with roster data and red-team notes to produce consolidated per-agent scores.

Automated performance reviews rely on predefined triggers like quarterly cadence or specific audit events, meaning ad-hoc evaluations outside these triggers require manual user requests to initiate the review process.