AI Asset Inventory

Qpoint is building an inventory page that shows the AI assets in use across an environment — agents, MCP servers, files, models/endpoints, and more. Before wireframing, we want a visual grammar grounded in what already works: what entity types incumbents show, whether they use tables, cards, facets, or graphs, what columns and posture overlays carry the story, and how "shadow" or unapproved assets get flagged.

1. Research plan → research-plan.md — target products by category and a screengrab capture method (c18-proven).
2. Precedent survey with screengrabs → precedents.md + precedents/ — per-product notes, downloaded screenshots with sourced README, extracted patterns, and implications for the eventual Qpoint inventory brief.

In (this pass — research only)

• research-plan.md — target list + capture strategy
• precedents/ — screengrabs with README.md source table
• precedents.md — synthesis: patterns, gap statement, implications
• Cycle bookkeeping (this file, CYCLES.md)

Out (next pass)

• Design brief (briefs/)
• app/pages/c19/ wireframe pages, fixture data, registry entries
• Mockups

• Cycle created and registered
• Research plan written (research-plan.md)
• Screengrabs collected (precedents/ + README) — round 1: 12 captures across 7 vendors
• Precedent synthesis written (precedents.md)
• Round 2: general (non-AI) inventory research — 6 more captures (NetBox, Backstage, Ralph, Tenable) + surveyed-without-capture notes
• Design brief (next pass)
• wf1 wireframes (next pass)