1---2name: first-touch-radar3description: >-4 Run the first-touch product radar for Design, Slides, and Clips. Use for daily5 watch, replay triage, weekly growth reports, or first-session incidents.6scope: dev7metadata:8 source: https://skillmd.com/skills/builderio/first-touch-radar9---1011# First-touch radar12131. Run `pnpm --filter analytics radar:watch --window daily` and read both the14 JSON and Markdown under `.tmp/first-touch-radar/runs/`.152. Treat exit 0 as clear, exit 1 as a fired, unavailable, or insufficient16 check needing attention, and exit 2 as a run that could not complete. Keep17 unavailable and insufficient metrics explicit; never turn missing evidence18 into zero or a pass.193. On a fire, run `pnpm --filter analytics radar:triage` and inspect the20 outcome-ranked sessions. Use `metric-watch` and `replay-triage` for details.214. For a weekly review, run `pnpm --filter analytics radar:report --week22 <Monday>` and read the coverage flags, denominators, campaign annotations,23 previous-issue recovery, and A/B maturity before drawing conclusions.24 The `campaign-followup` check runs in daily watch and weekly report; compare25 identified recipients' post-send activity with the seven days before each26 send, split by recipient app and activity app. Weekly reports cap cumulative27 BigQuery dry-run estimates at 100 GB; use `--max-total-gb` to set a different28 report-wide cap, and treat a reached cap as incomplete evidence.295. Run `feedback-ingest` for Slack themes. Searches are read-only and begin at30 the saved watermark.316. Send one Telegram to Steve only when a check fired, is unavailable or is32 insufficient, or the weekly report is ready. Include counts and the private33 run-file path. Send nothing on a quiet day. Never send Slack or email.347. Keep recipient emails, HMAC keys, and prompt text out of logs, reports,35 Telegram, and commits. The campaign check reads its mode-0600 recipient file36 in memory, binds keyed email digests as query parameters, disables local and37 BigQuery result caching, and emits aggregate counts only. BigQuery matching38 requires `ANALYTICS_RADAR_CAMPAIGN_HMAC_KEY` and a configured39 `ANALYTICS_RADAR_CAMPAIGN_HMAC_FUNCTION` backed by the same secret server-side.40 The function accepts a normalized email and returns lowercase hex41 `HMAC_SHA256`; keep its key in Secret Manager and never pass the key to42 BigQuery. When either setting is missing, keep the campaign check43 unavailable. Unknown signup campaign labels are grouped as `other`.44 Anonymous landings cannot be assigned to a recipient; do not interpret45 identified counts as a complete click total.468. This radar changes scripts and skills only. Do not make user-visible UI47 changes. If UI work becomes necessary, stop that scope and request Steve's48 screenshot approval before implementing it.499. A fired metric may create a local brief through `fix-brief`; no fix starts50 until Steve approves it.5110. Interpret the first-touch journeys separately: Design web excludes52 `/mcp/oauth%` signup paths; Design MCP is signup → `agent_connected` → first53 `$mcp_tool_call` → `design_created`. Clips activation is signup →54 `recording_ready` within 24 hours. Treat 2026-10-21 as a definition break:55 `recording_started` before then includes local-only recordings.5611. Read the weekly paid-acquisition notes alongside signup counts: Slides ads57 stopped after 2026-10-16, Bing CPC ended around 2026-10-20 21:00 UTC for58 Clips/Design, and Google paid spend was about 45% lower from 2026-10-21.