docs: \"Operate your Trinity in 5 minutes\" onboarding quickstart for the ops starter

[vybe]

Summary

Write a getting-started quickstart that takes a first-time operator from zero to a running ops/DevOps starter agent in ~5 minutes, with copy-paste steps and a set of sample ops prompts ("check fleet health", "show me recent errors", "update and roll back if it fails").

Context

The starter agents only drive adoption if the path to using them is obvious. A focused quickstart + prompt cookbook is the front door for new operators and the reference the deploy flows link to.

Acceptance Criteria

• Quickstart under docs/user-docs/ covering connect → deploy ops starter → first ops command
• Sample ops prompt cookbook (health, logs, updates, rollback, provisioning)
• Uses placeholder values only (your-domain.com, ${MCP_API_KEY}, user@example.com) — no real hosts, emails, or secrets
• Linked from the relevant deploy/onboarding entry points

Technical Notes

• Cross-link abilityai/trinity-ops-public and the abilities trinity plugin skills (/trinity:connect, /trinity:onboard, /trinity:deploy-new-instance).

Parent epic: ops/DevOps starter agents for one-click instance management.

[oleksandr-korin]

Partial audit — the surface has shifted since this was filed.

What has shipped since 2026-06-15

• docs/user-docs/abilities/trinity-plugin.md (241 lines) — documents the current 7 skills (this ticket knew 3), including the newer /trinity:start-here, /trinity:loop, /trinity:sync, /trinity:create-dashboard
• /trinity:start-here — landed via docs: document /trinity:start-here, the guided first journey #2054 (2026-08-15). A resumable, interactive walkthrough covering "what is Trinity → get an instance → connect → first agent alive." Substantially the "5-minute quickstart" this ticket asked for, in interactive command form
• docs/user-docs/operations/ now has audit / dashboard / executions / monitoring / operating-room / telemetry pages

What is still missing (the real gap)

1. An ops-starter-specific walkthrough. Today's docs describe the plugin's skills generically. There is no dedicated end-to-end "spin up the DevOps starter agent that manages this Trinity instance, from empty to first useful command" narrative. /trinity:deploy-new-instance is documented but not demonstrated with concrete outputs.
2. The prompt cookbook. The ticket's second artifact. Nothing like it exists in docs/user-docs/. This is where new value lives — a curated list of prompts that produce useful output from an ops starter: fleet health, error triage, updates, rollbacks, provisioning.
3. Cross-links from docs/user-docs/operations/*.md to the ops-starter workflow.

Suggested narrowed scope

Update AC to:

• docs/user-docs/operations/ops-starter-quickstart.md — end-to-end ops-starter deployment + first command, verified against a real running instance (not written from inference — the Docs advertise two install paths that fail on a fresh machine (install.sh one-liner, quickstart.sh) — repoint to start.sh #1788 lesson)
• docs/user-docs/operations/ops-prompt-cookbook.md — the curated prompt library (health / logs / updates / rollback / provisioning)
• Cross-links from operations/{monitoring,operating-room,executions}.md and from abilities/trinity-plugin.md § /trinity:deploy-new-instance
• Placeholder discipline as originally specified (your-domain.com, ${MCP_API_KEY}, user@example.com)

At the narrowed scope this is closer to 60–90 min of writing plus the real-instance verification. Priority (P3, complexity-low) still tracks.

Not closing — flagging that whoever picks this up should treat the "5-minute quickstart" half as substantially delivered by /trinity:start-here, and focus on the two missing artifacts above.