Home / Home / Scenarios / Scenario: Cross-Agent Budget Race

Scenario: Cross-Agent Budget Race

Two agents share a daily budget. One spends early and fast; the other starves — or both loop retrying. The shared ceiling needs a design, not just a number.

What happens

Agent A consumes the shared ceiling by lunch; Agent B's legitimate work gets BLOCKED all afternoon.

Teams respond by raising the ceiling — which defeats the budget.

What happens with sipi.bot

Per-agent ceilings on top of the shared ceiling: each agent gets its own slice AND the fleet has a cap.

Priority handling: critical agents get higher ceilings; the shared cap stays intact.

The rule set

Shared fleet ceiling + per-agent ceilings + approval threshold for over-budget requests.

The design

Fleet ceiling: $1,000/day
  Agent A: $600/day (critical path)
  Agent B: $400/day (batch)
Over-budget requests: FLAGGED for approval

Slices + a cap beats a race.

FAQ

Is a shared ceiling bad?

No — it's necessary. The fix is per-agent slices inside it.

What if a critical agent needs more mid-day?

The approval queue handles it: FLAG over-budget requests, approve the ones that matter.

Stop the next $12,400 night.

One API call (or MCP tool) in front of every agent transaction — APPROVED, BLOCKED, or FLAGGED, deterministic, ~5 ms, fully logged.

See plans — from $99/mo Try a live check