Pume command deck
A truth-first cockpit pass for internal visibility across the website, agents, ideas, costs, infrastructure, and workflow telemetry.
Static exportSynced 20 MarTruth labels added2 epics linked
Operating picture
A calmer shell, now with clearer truth about what is real and what is still missing.
The correction path starts here: real ledger cost, real capability scoring, real repo parity, and explicit signals when a surface is still only partially wired.
4 real source lanesTruth labels in progressStatic export friendly
74%Readiness
The shell is strong, but depth now depends on replacing remaining seeded surfaces with live or generated evidence.
Pulse
Real + explicitThis week
Runs30d
Agentslive
Epicslinked
Cost$30d
Syncrepos
Linked delivery work2 epics / 8 tasks
10PRDA-74 is already translated into real delivery work, not just a parked concept.
Lab runs in 30 days24 Feb
397The cockpit now inherits a real signal lane from the same lab snapshot used for public-safe reporting.
Tracked AI spend23d old
$7.01This is canonical run-ledger spend for the last 30 days, not a budget placeholder.
Major-work eligible agents1h old
3Capability and eligibility now come from the generated agent capability report.
Focus
Four prioritiesActive tracks
Ledger-backed cost lane100%
Agent capability wiring78%
Repo drift surface66%
n8n runtime API28%
Recent movement
Latest truthful eventsOperational notes
Pume Cockpit v1active
The cockpit experiment is now visible in the public-safe lab stream and anchored in the mature pack.
Delivery bundle createdlinked
2 epics and 8 tasks were staged directly from PRDA-74.
Lab snapshot synced into the shellfresh
Current feed reports 397 runs in 30 days and 20 active agents.
Cost lane is explicit about coveragetruth-first
The cockpit now calls out that tracked spend is AI-run ledger data and that hosting/tool invoices still need adapters.