GTM Engine — mission control

GTM Engine mission control

▾ {{ campId }}{{ campChip }}

parked Jul 2 · ISC-17 · creds stripped to env: refs

domain GREEN

warmup day 1 · sent {{ sentToday }} · cap 40/day

daemon · step {{ stepNum }}/18 — {{ stepName }}

nothing sends itself — the engine writes, you approve, Courier fires

Parked

campaign · dormant since Jul 2 · ISC-17

target roles

product spec on disk: {{ parkedInfo.product }}

To restore: move this block from local/_parked-campaigns.json back into config.json → campaigns[], fill the env: creds — the daemon picks it up next tick. Same belt, same gates, different product.

← back to sg-vets

Morning, operator. {{ qCount }} drafts want your OK.

The engine writes, you approve, Courier fires. Watch a lead run the line →

{{ pl.big }} {{ pl.small }}

Daemon heartbeat

↻ loops forever · 18 steps · click one

PAUSED

FIXED

Approval queue

oldest first

Review one by one →

“{{ tq.subject }}”

→ {{ tq.company }}

{{ tq.wc }}w

Activity

Drift · 5 open

The Line — every manipulation, one by one

replaying Aldervale Animal Specialist & Emergency Hospital · the engine's one fully recorded run (Apr 21) · every number, document and verdict below is real

station {{ lineNum }}/15

⏮

‹

›

On the belt — in · before

+ {{ ti }}

→

{{ trace.dur }}

what it does — found · measured · decided

{{ trace.src }}

Off the belt — out · + new lines highlighted

{{ trace.outMono }}

📄 read the full artifact →

Station 15 · the aftermath

Sent at 06:18. Quarantined by 07:00.

This run is why the belt now stops at station “You”. No draft reaches Courier without a human verdict — the {{ qCount }} letters in your approval queue are the direct fix.

Review the {{ qCount }} waiting →

cost of the lesson · $0.947 · 141.5k→39.8k tokens

The lead's file · builds as it rides

spend so far{{ lineSpend }}

tokens{{ lineTokens }}

status{{ lineStatus }}

To: {{ deskCard.company }}

sg-vets · {{ deskCard.wc }}w

“{{ deskCard.subject }}”

{{ deskCard.body }}

Also queued — follow-up, day +5 · {{ deskCard.fuwc }}w

{{ deskCard.fu }}

✦

Desk clear.

{{ sentToday }} approved → Courier · {{ heldCount }} held → needs-review
Follow-ups self-schedule for day +5. See you at the next tick.

↺ Reset demo

The desk

every draft the engine wrote, one decision at a time

reviewed {{ reviewedCount }} / 13

✦ {{ sentToday }} approved↩ {{ heldCount }} held

APPROVE ✦

HOLD ↩ needs-review

‹ back

skip ›

What approve does

Hands the letter to Courier inside today's budget (cap 40, window 09–17 recipient-tz). Follow-up self-schedules day +5. Hold parks it as needs-review.

why this desk exists ⚠ the one time the engine sent without a human, it cost $0.95 and a quarantine — watch the replay

Leads

campaign sg-vets · showing 28 of 91 · click a row for its journey

Company
Campaign
Status
Emails
Draft

sg-vets

{{ lr.emails }}

full census: bun run reports · 91 leads on disk

Drift — code vs. intent

from the July 2026 re-audit · 8 findings fixed since v1, 5 still open · FIX lines are engineering proposals, not audit findings

⚠ Still open · 5

IF IGNORED — {{ dr.ignored }}

FIX — {{ dr.fix }}

{{ dr.where }}

✓ Fixed since v1 · 8

FIXED

{{ df.ttl }} — {{ df.what }}

ArtifactPackage · what Artisan actually built

WhatsApp Automation Audit — Aldervale

209.6s · $0.296 · 64,238 → 9,361 tokens · local/artifacts/aldervale-…/artifact.md · verbatim

Critic · PASS Judge · PASS ISC · 1 violation

⚠ ISC-NEVER-ASSUMPTIVE-PAIN — this pain was labeled “hypothesis-validated”, but the evidence has no transcript or call reference. Logged 06:12:19; log-only in April.

This is the “full audit” the email offers to send — built before the first email, so the promise is real on day one.

Lead · sg-vets

{{ selLead.status }} emails {{ selLead.emails }}

▶ Replay on the Line

📄 view its artifact

Journey · click a stage

{{ stageDetail.io }}

☑{{ ck }}

↳{{ ex }}

⚠ {{ stageDetail.drift }}

⚠ Constraint violation

{{ selLead.violation.id }}

Status history · real

Stage spend · real