Journal topic
Guides.
How to cap, approve, and record what AI agents do: concrete setups with commands you can run.
-
Claude Code spend limit per developer: what it caps
How the Claude apps gateway caps spend per developer, group or org, what it returns at the limit, where it fails open, and what a per-developer cap leaves open.
-
Per-agent API keys or per-agent budgets: what each stops
A key names the caller, a budget counts spend. How OpenAI projects and OpenRouter keys attach caps, and where a per-agent key still leaves spend uncapped.
-
AI agent kill switch: what it has to revoke
A kill switch for an AI agent is revocation, not a process stop. What to revoke at the key, token, card and status-list layers, and what stays out of reach.
-
How to stop an AI agent infinite loop: four layers
Turn caps, SDK budgets, rate limits and an outside spend cap each stop a runaway agent loop differently. What each counts, returns and misses.
-
LiteLLM budget per key: what it enforces and where it ends
How LiteLLM key budgets work: max_budget, resets, pre-call reservation, three documented ways a budget stops capping, and who can change it. Read 14 October 2026.
-
Cloudflare AI Gateway spend limits: what they enforce
What Cloudflare AI Gateway spend limits check, what a 429 means, where a burst can pass the cap, and what a per-agent budget rests on. Read 10 October 2026.
-
Virtual card limits for an AI agent: what Stripe enforces
How to cap an AI agent's virtual card with Stripe Issuing: spending controls, real-time authorizations, their gaps, and one cap shared with LLM spend.
-
Stolen AI agent API key: what proof of possession fixes
What a thief can do with a leaked agent key, what short lifetimes and proof of possession change, and what they leave open. With RFC quotes and a demo.
-
How to set a daily spending limit for an AI agent
Anthropic and OpenAI cap spend by the month. Where a daily window exists, what each cap returns when it trips, and how to stop an agent retrying into it.
-
AI agent approval fatigue: a signed limit beats prompts
Why per-action approval prompts stop working, what auto mode and allow rules trade away, and how a held request with a threshold asks a human when it matters.