gtm-brain

The agent-value review. Evaluates how much value the workspace is actually getting out of its Day AI agents — who's in and who should be (with draft nudge emails), whether everyone has the agents they should (≥2 each; the right archetypes), and whether each agent's skills and identity are strong. Produces a scored report and a prioritized, ready-to-execute recommendation set. Recommend-only. Usage: /agent-audit [a person, team, or dimension to focus on]

The entrypoint for GTM Brain. Verifies the Day AI MCP connection and the operator's role, takes stock of every initiative in initiatives/, reports progress against each one's verifiable success criteria, and kicks off the agents and skills the active initiatives need. On a fresh clone it runs the default bootstrap initiative (connect, identify people, scaffold the plan). Run this whenever you sit down to work. Usage: /start [an initiative slug or focus]

Vibe-code a custom app or integration on top of Day AI using the public Day AI SDK (github.com/day-ai/day-ai-sdk) — internal tools, dashboards, mashups, automated workflows, anything a skill or Page can't express. Knows what the SDK offers, how to get started, and which example template to clone. Usage: /build-app [what you want to build]

The teaching guide for authoring a single high-quality Day AI skill prompt. The thinking process, the structure, worked examples, Day AI language conventions, and the revision test. Read this before writing any skill. Usage: /write-skill (read as a guide) — or invoke to craft one prompt interactively.

Sync the planning documents and active initiatives to/from Day AI Pages so the rest of the company sees the same source of truth. Pushes local planning/*.md and initiatives/*.md to Pages, or pulls Pages back into the repo. Degrades gracefully if Pages tools aren't available on the connected MCP. Usage: /sync-pages [push|pull|status]

Compare the current Day AI workspace (members, agents, skills) against the planning layer and produce a prioritized gap report. Read-only — makes no changes. Run before /implement. Usage: /audit [a specific person, team, or outcome to focus on]

Design one complete, deployment-ready Day AI agent for a person — its identity (name, title, description = its system prompt) and its 1–3 starter skills — built on a proven archetype (CRM Data Nerd, Coach, Relationship Radar, etc.) and grounded in the person's real role and the company's process. The companion to /write-skill (one skill) — this designs a whole human-shaped agent. Usage: /design-agent {person} [archetype or job-slice]

Turn the plan and audit into real configuration in Day AI — invite the right people at the right roles, tune each teammate's agent identity, and deploy role-specific skills. Always previews before it writes. Requires Owner/Admin. Usage: /implement [scope: a person, team, outcome, or "the high-priority audit items"]

Build or refresh the planning layer — the company plan, the CRO-level strategy, the concrete outcomes, and the who's-who. Runs the gtm-strategist through an interview loop and grounds the plan in the workspace graph. Usage: /plan [layer-1|layer-2|layer-3|people | a free-form focus]