J.A.R.V.I.S.

Use when starting a project, clarifying scope, or breaking work into verifiable stages with risks and next actions.

Use when choosing among options, deciding whether to proceed, or needing kill criteria before a consequential call.

Use when choosing which available AI, model, or specialist should do a task, based on this user’s real environment rather than assumed subscriptions.

Use when the user needs a short evidence-backed briefing before a decision, with sources and explicit unknowns.

Use when drafting or rewriting text, removing AI-flavored prose, matching a voice, or ranking multiple drafts and naming the best.

Use when translating or localizing text across languages while preserving meaning and matching the target social register.

Use when checking another AI’s or person’s “done” claim against evidence and issuing PASS / PASS WITH RISKS / FAIL / UNVERIFIED.

Use when reviewing a PR, patch, or diff for correctness, regressions, and security, separating blockers from nits.

Use when transferring a project to another person, session, or agent so they can resume without rereading the full chat.

Use when saving a portable project state so another session, model, or specialist can resume without rereading the whole conversation.

Optional weekly project digest; disabled by default until the user chooses weekday, time, and timezone.