← Back to brief

project adam status 2026-05-13

memory · project_adam_status_2026-05-13.md

5/12 18:46Z — Adam's executive-approval gate (verbatim)

Adam Fleckles, replying to Joseph's brain-dump send:

> "Before I share full login credentials for the systems I use, we need to connect with the Executive team to get approval here. With that said, let's keep this simple and move forward with what we have for now. It will likely take some time to provide you with full authorization for all these internal systems."

What this means in plain terms:

Phase 1A scope — locked 2026-05-13

WorkstreamPhase 1A status
Adam's PWA shell✅ Stays in lock-step with GC system (mirror-deploy hook in deploy_pwa.sh step 10/10)
Adam's tenant runtime✅ Provisioned + dormant (per project_adam_tenant_dormant.md)
Adam Google Workspace OAuth⏸ Was scheduled Thursday 5/14 15-min screenshare — now blocked on exec approval
Adam HubSpot owner_id⏸ Was PENDING_FROM_ADAM in tenant.json — still pending, no urgency this week
Adam Reamze / Notion / Data Studio / MARQ / Fathom credentials⏸ ALL blocked on exec approval
Wednesday 5/13 stand-up✅ Still happening — agenda becomes "what we've built in parallel + what's ready when access arrives"

Joseph's operator directive (2026-05-13)

Verbatim instruction received this session:

> "Push forward on the GC system; keep Adam's PWA in lock-step with the GC model; don't ask Adam questions; build for when access does come."

Operational rules this drives:
1. No questions to Adam about credentials, owner_id, OAuth, or anything that blocks on Exec approval. We wait.
2. GC system continues as primary — all skill / cron / PWA / outreach work on Joseph's GC instance proceeds as normal.
3. Adam PWA shell mirrors GC PWA — every time deploy_pwa.sh runs, Adam's PWA gets the same shell update automatically (Deliverable 2 of this turn wires the mirror step).
4. No Adam data files are deployed (no leads.json, deals.json, outbound.json, approvals.json, source_health.json from Joseph leaking to adam-bd-brief.pages.dev). Mirror copies HTML/CSS/JS shell only.
5. Adam-specific files stay Adam-specific — Adam's own index.html (redirect-to-/preview shell), _worker.js (adam_bd_approvals KV binding), manifest stay intact.

Wednesday 5/13 stand-up — revised agenda

Since "instance stood up" is off the table, the call posture shifts to:

What's already done that doesn't need Adam's creds

These keep running on Joseph's instance and continue to harden the system Adam's tenant will eventually receive:

What blocks (and unblocks how)

Blocked workstreamUnblocks when
Reamze SMS ingestionAdam delivers Reamze API key or grants chrome_bridge profile session
Fathom transcript ingestionAdam grants Gmail OAuth scope (from:noreply@fathom.video filter already wired into transcript-scan) OR Fathom API token
HubSpot owner_id filterAdam pastes the /users/<owner_id> URL from his HS profile (he can do this without exec approval — it's his own URL, not a credential)
SkyRun Data Studio new-lead intakeAdam delivers Looker Studio export-share OR the underlying Google Sheet ID + read OAuth
Notion notes importAdam shares the workspace via integration token
MARQ asset inventoryAdam shares the workspace or exports a manifest
QRCode Tracker (SPMS)Adam shares vendor portal access OR an export schedule
Dimaco Mailing TrackerAdam shares vendor portal access OR a CSV export
OwnerPoint (legacy)Adam confirms credentials still valid + grants access
Vintory (legacy)Adam confirms credentials still valid + grants access

Cross-references

How to use this memory

When the next Adam contact happens (Wednesday 5/13 stand-up, or any inbound from Adam):
1. Re-read this file first — operator posture is "don't push, parallel-build, mirror-deploy"
2. Cross-check reference_adam_tool_inventory_2026-05-13.md for the per-tool plan
3. Do NOT cite specifics from this synthesis in outbound to anyone outside SkyRun without source-verifying against the Gmail thread (Adam's 5/12 18:46Z reply)
4. Update this file (not predecessor) with delta from the next Adam touchpoint