Quality-only antislop review of changed code (or a given scope) via the rival binary — hunts slop and over-engineering, returns a leanness rating and a cut list, never bugs. Default model Sol at xhigh effort. Detached + watched in the background. Use only…
Code review via Fable through the rival binary — reviews changed files (or a given scope) at its configured effort. Detached + watched in the background. Use only when the user explicitly invokes /rival-fable.
Run Grok (grok-4.6) through the rival binary, detached and watched in the background. Use only when the user explicitly invokes /rival-grok.
Run Kimi K3 (max reasoning, via opencode) through the rival binary, detached and watched in the background. Use only when the user explicitly invokes /rival-k3.
Review a plan/spec markdown document with Fable only via the rival binary. Rates it 1-10 and finds bugs and gaps. Use only when the user explicitly invokes /rival-plan-fable.
Review a plan/spec markdown document with Sol and Fable in parallel at xhigh effort via the rival binary. Each rates it 1-10 and finds bugs and gaps; both results are shown. Use only when the user explicitly invokes /rival-plan.
Review a plan/spec markdown document with Sol at xhigh effort via the rival binary. Rates it 1-10 and finds bugs and gaps. Use only when the user explicitly invokes /rival-plan-sol.
Run Sol, K3, and/or opt-in grok code reviews with a consilium judge via the rival binary. Use only when the user explicitly invokes /rival-review.