AI & Swarm OperationsSuccess Rate
12 days
Agent RegistryPerformanceSuccess RateCostCollaborationFailuresAutonomous ActionsSwarm Health

Agent Success Rate

Per-discipline run outcomes across the build crew, ranked highest first. ● live from agent_runs

Swarm-wide

Weighted by run count, so heavy-traffic disciplines move the needle more than low-volume ones.

Weighted success rate
93.9%
Across 61 runs
Total runs
61
All disciplines
Total failures
9
At / above 90%
9/ 10
Reviewer / Drift top, Realtime lowest

Success rate by agent

Stacked bar showing percentage of successful runs; reference line at 90% target.

1007550250Reviewer / DriftValidation EngineDev-Admin CockpitSchemaWorkerQuality / TestAPI / Write-PathUI KitInfra / DevOpsRealtimetarget

Ranked by success rate

Bar width tracks the success rate; color follows the threshold (≥90% green, ≥75% amber, else red).

  • Reviewer / Drift
    3 runs100%
  • Validation Engine
    6 runs98%
  • Dev-Admin Cockpit
    4 runs97%
  • Schema
    9 runs · 1 fail96%
  • Worker
    4 runs95%
  • Quality / Test
    9 runs · 1 fail94%
  • API / Write-Path
    7 runs · 1 fail93%
  • UI Kit
    8 runs · 2 fail91%
  • Infra / DevOps
    6 runs · 2 fail90%
  • Realtime
    5 runs · 2 fail88%

Threshold legend

The same traffic-light bands used across the cockpit.

≥ 90% — healthy75–89% — watch< 75% — at risk