Cycles

Appearance

The Cycles Blog ​

Engineering insights, product updates, and best practices for AI agent runtime authority, exposure, and cost control.

Tools you can run on your own numbers: Cost Calculator (Claude vs GPT) → · Blast Radius Risk Calculator → — every configuration produces a shareable URL. Many of the posts below link to a calculator pre-loaded with the post's specific scenario.

RSS

The Model Call Succeeded. Then the Process Died.FEATURED

· Albert Mavashev · 5 min read

How Cycles preserves known AI agent spend across lost responses, expired reservations, credential failures, and process restarts.

Cursor Agent Deleted a Prod Database in 9 SecondsFEATURED

· Albert Mavashev · 9 min read

Cursor agent reportedly wiped a Railway production database and backups in one API call. Prompts aren't permissions — agents need pre-execution gates.

Agents Are Cross-Cutting. Your Controls Aren't.FEATURED

· Albert Mavashev · 9 min read

Agents span providers, tools, tenants, and workers. Learn why governance needs cross-cutting budget authority plus application-side authorization at runtime.

State of AI Agent Governance 2026FEATURED

· Albert Mavashev · 14 min read

AI agents moved to production faster than governance kept up. This is the state of enforcement, regulation, and incidents in 2026 — and what comes next.

AI Agent Risk Assessment: Score, Classify, EnforceFEATURED

· Albert Mavashev · 17 min read

Classify AI agent actions by blast radius, assign consistent risk scores, and translate each assessment into enforceable runtime budget controls in practice.

Portable Agent Skills Need Runtime Budget TestsNEW

· Albert Mavashev · 8 min read

Test agent skill budgets across CLI and cloud runtimes. Use a shared task ledger, concurrent admission checks, and a runnable Cycles budget portability drill.

When AI Agents Rewrite Approval Rules

· Albert Mavashev · 7 min read

AI agents can configure the approval rules that govern future actions. Learn how to separate drafting from publishing, test exceptions, and verify changes.

When an AI Agent Works Around a Denial

· Albert Mavashev · 8 min read

AI agents can retry blocked actions through other tools or workers. Learn how to preserve policy boundaries, permit safe fallbacks, and test what happens next.

AI Agent Rollbacks Can Make Incidents Worse

· Albert Mavashev · 8 min read

AI agent rollbacks can overwrite legitimate changes. Learn how scoped recovery permissions, version checks, and replay protection limit damage during repairs.

LLM Failover Needs a Shared Budget

· Albert Mavashev · 7 min read

LLM failover can spend twice on one task. Learn how shared budgets, per-attempt reservations, and usage reconciliation keep provider fallback costs accountable.

Prompt Caching Costs and AI Agent Budgets

· Albert Mavashev · 13 min read

Prompt caching gives one agent call three prices. Learn why AI agent budgets should reserve for the cache write, settle from usage, and track cache read share.

Agentic RAG Cost Control: Budget Every Answer

· Albert Mavashev · 14 min read

Control agentic RAG costs across retrieval, reranking, and model calls. Learn to enforce per-answer budgets with Cycles and handle budget exhaustion safely.

You Stopped the Agent. Its Jobs Kept Running.

· Albert Mavashev · 8 min read

Stopping an AI agent can leave queued tasks and provider jobs running. Build a shutdown procedure to block new work, verify cancellation, and settle usage.

Pace AI Agent Authority as Models Advance

· Albert Mavashev · 6 min read

Dario Amodei calls for pacing frontier AI. For production teams, runtime authority offers a practical way to bound agent spending, actions, and delegation.

Grok Bot Has a Computer. Where Does Cycles Fit?

· Albert Mavashev · 11 min read

Grok Bot gives persistent AI teammates a shared cloud computer. See where Cycles can enforce budgets and action authority—and where it cannot.