Technical articles about AI agent infrastructure — durable execution, observability, and developer tools.https://blog.axme.ai/en-ushttps://blog.axme.ai/after-100-sessions/https://blog.axme.ai/after-100-sessions/Before: re-explain the project every session, watch the same mistakes, fix the same bugs twice. After 100 sessions with persistent memory: different patterns, measurable time savings, one problem I did not expect.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsproductivityhttps://blog.axme.ai/claude-md-is-read-only/https://blog.axme.ai/claude-md-is-read-only/Everyone puts project context in CLAUDE.md. It's static, it grows, nobody maintains it, and it becomes a graveyard of stale rules. Writable memory with an audit trail is a different shape of solution.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsmemoryhttps://blog.axme.ai/conversation-with-claude-code-47-times/https://blog.axme.ai/conversation-with-claude-code-47-times/Every session starts the same way: explaining your stack, your deploy flow, your conventions. You've said it forty times. The agent keeps nodding. Here's why re-explaining every session is the wrong fix.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsproductivityhttps://blog.axme.ai/decisions-are-not-memories/https://blog.axme.ai/decisions-are-not-memories/'I learned we use Postgres' and 'we decided Postgres over MySQL because of JSON support' are different data. One is a fact. The other is a contract. Mixing them is why agent memory drifts.Sat, 11 Apr 2026 00:00:00 GMTcodeaiagentsmemoryarchitecturehttps://blog.axme.ai/first-session-new-codebase/https://blog.axme.ai/first-session-new-codebase/Stack detection, structure scanning, glossary extraction, safety rule seeding. What happens in the first 30 seconds of a new AXME Code setup, why each step matters, and what happens if you skip it.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsonboardinghttps://blog.axme.ai/four-types-of-context/https://blog.axme.ai/four-types-of-context/Most agents treat all context as 'stuff in the context window'. But facts, patterns, project state, and references have different lifecycles and different retrieval rules. Mixing them is why agents feel dumb.Sat, 11 Apr 2026 00:00:00 GMTcodeaiagentsmemorycontexthttps://blog.axme.ai/multi-repo-workspaces-break-claude-code/https://blog.axme.ai/multi-repo-workspaces-break-claude-code/You have 5 repos in one workspace. Each has its own conventions. A single CLAUDE.md is the wrong shape. Per-repo storage with a workspace root is the pattern that actually scales.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsmonorepohttps://blog.axme.ai/prompts-dont-stop-force-push/https://blog.axme.ai/prompts-dont-stop-force-push/You put 'never force push' in CLAUDE.md. The agent force-pushed anyway. Here's why prompt-level safety rules fail and what tool-level enforcement actually looks like.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolssafetyhttps://blog.axme.ai/stop-writing-claude-md/https://blog.axme.ai/stop-writing-claude-md/CLAUDE.md is a passive artifact. Decisions are active rules with enforcement levels, supersede chains, and audit history. Here's why I stopped maintaining CLAUDE.md and what I switched to.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsworkflowhttps://blog.axme.ai/subagents-lose-context/https://blog.axme.ai/subagents-lose-context/When Claude Code spawns a subagent to run a task in parallel, the subagent starts fresh. It doesn't inherit the parent's memory. Work gets done, but context is lost. Here's the fix, and why it's fragile.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolssubagentshttps://blog.axme.ai/vacation-empty-project-session-handoff/https://blog.axme.ai/vacation-empty-project-session-handoff/Five days off. Session handoff gone. Half a week of decisions forgotten. Here's the 'what was I doing yesterday' problem and why session close is a ritual that matters.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsworkflowhttps://blog.axme.ai/we-built-axme-code-with-axme-code/https://blog.axme.ai/we-built-axme-code-with-axme-code/Six months of dogfooding. Real memories, real decisions, real safety rules from our own repo. What patterns emerged. What we threw away. What surprised us.Sat, 11 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsdogfoodhttps://blog.axme.ai/ai-agent-crash-recovery-default/https://blog.axme.ai/ai-agent-crash-recovery-default/Your agent ran 47 steps of a 50-step pipeline. Then it crashed. The state is gone. Every framework says you should have configured checkpointing. Why isn't durability built in?Wed, 08 Apr 2026 00:00:00 GMTcloudaipythonagentsdurabilityhttps://blog.axme.ai/claude-code-amnesia/https://blog.axme.ai/claude-code-amnesia/Claude Code forgets your project every session. CLAUDE.md helps, but it doesn't scale. I built an MCP server plugin that gives Claude Code persistent memory, decisions, and safety guardrails.Wed, 08 Apr 2026 00:00:00 GMTcodeclaudeaidevtoolsproductivityhttps://blog.axme.ai/human-approval-ai-agent-workflow/https://blog.axme.ai/human-approval-ai-agent-workflow/Adding a human approval step to an AI agent workflow means building a notification service, reminder scheduler, escalation chain, and webhook handler. Or you can use 4 lines of code.Wed, 08 Apr 2026 00:00:00 GMTcloudaipythonagentshumanintheloophttps://blog.axme.ai/stop-rogue-ai-agent-production/https://blog.axme.ai/stop-rogue-ai-agent-production/Your AI agent went rogue at 3am. It's running on multiple instances across regions. There's no terminal to Ctrl+C. You need a kill switch that works in under 1 second, enforced at the infrastructure level.Wed, 08 Apr 2026 00:00:00 GMTmeshaipythonagentsmonitoringhttps://blog.axme.ai/temporal-alternative-no-cluster/https://blog.axme.ai/temporal-alternative-no-cluster/Temporal is the go-to for durable execution. But it requires a cluster and forces determinism constraints on your workflow code. Here's an alternative that gives you durability through a managed API with none of those trade-offs.Wed, 08 Apr 2026 00:00:00 GMTcloudtemporalpythondurableexecutionworkflowhttps://blog.axme.ai/ai-agent-fleet-dashboard/https://blog.axme.ai/ai-agent-fleet-dashboard/AI agents are multiplying across your org. Different clouds, different frameworks, different teams. Without a fleet dashboard, you're flying blind.Sun, 05 Apr 2026 00:00:00 GMTmeshaidashboardmonitoringagentshttps://blog.axme.ai/ai-agent-cost-monitoring/https://blog.axme.ai/ai-agent-cost-monitoring/AI agents call LLMs. LLMs cost money per token. Nobody tracks it per agent. One runaway loop and your OpenAI bill is a disaster.Thu, 02 Apr 2026 00:00:00 GMTmeshaipythonllmmonitoringhttps://blog.axme.ai/ai-agent-governance-platform/https://blog.axme.ai/ai-agent-governance-platform/Most teams can't tell you which agents are running, how much they've spent, or how to shut one down. Here's what a governance platform for AI agents looks like.Thu, 02 Apr 2026 00:00:00 GMTmeshaigovernanceagentsenterprisehttps://blog.axme.ai/ai-agent-health-monitoring/https://blog.axme.ai/ai-agent-health-monitoring/Your agents are running across 4 machines. One dies. No alert. No log. You find out 3 hours later from a customer complaint. Here's how to fix that.Thu, 02 Apr 2026 00:00:00 GMTmeshaipythonmonitoringdevopshttps://blog.axme.ai/ai-agent-heartbeat-monitoring/https://blog.axme.ai/ai-agent-heartbeat-monitoring/Container is green. Process is running. But your agent stopped processing work 2 hours ago. Heartbeat monitoring catches what health checks miss.Thu, 02 Apr 2026 00:00:00 GMTmeshaimonitoringpythonagentshttps://blog.axme.ai/ai-agent-kill-switch/https://blog.axme.ai/ai-agent-kill-switch/AI agents go rogue. They send thousands of emails, make unauthorized API calls, burn through budgets. You need a kill switch that works in under 1 second.Thu, 02 Apr 2026 00:00:00 GMTmeshaipythonagentsmonitoringhttps://blog.axme.ai/ai-agent-policy-enforcement/https://blog.axme.ai/ai-agent-policy-enforcement/AI agents ignore restrictions in prompts. Gateway-enforced action policies block unauthorized actions before they reach the agent - no prompt injection can bypass it.Thu, 02 Apr 2026 00:00:00 GMTmeshaisecurityagentsgovernancehttps://blog.axme.ai/ai-agent-rate-limiting/https://blog.axme.ai/ai-agent-rate-limiting/One runaway retry loop. 10,000 API calls. $130 in LLM costs. No rate limit fired because you never built one. Here's how to add centralized rate and cost limiting to AI agents.Thu, 02 Apr 2026 00:00:00 GMTmeshairatelimitingpythonagentshttps://blog.axme.ai/a2a-without-durable-lifecycle/https://blog.axme.ai/a2a-without-durable-lifecycle/Google's A2A protocol handles agent communication. But crash recovery, retries, timeouts, and human approval gates? That's still on you. Unless you add a lifecycle layer.Sun, 29 Mar 2026 00:00:00 GMTcloudaipythonagentsa2ahttps://blog.axme.ai/agent-timeout-escalation/https://blog.axme.ai/agent-timeout-escalation/API call hangs. Agent blocks. No timeout fires. No one gets paged. The default behavior in most agent frameworks is silence. Here's how to fix that.Sun, 29 Mar 2026 00:00:00 GMTcloudaipythonagentsreliabilityhttps://blog.axme.ai/approval-chain-without-workflow-engine/https://blog.axme.ai/approval-chain-without-workflow-engine/Manager approves, then finance approves. Simple enough to describe. 300 lines of code to build. Unless your platform handles approval chains natively.Sun, 29 Mar 2026 00:00:00 GMTcloudworkflowpythonbackendarchitecturehttps://blog.axme.ai/autogen-agents-different-machines/https://blog.axme.ai/autogen-agents-different-machines/AutoGen handles multi-agent conversations beautifully - inside one process. Put agents on different machines and you're back to building message brokers from scratch.Sun, 29 Mar 2026 00:00:00 GMTcloudaipythonagentsautogenhttps://blog.axme.ai/stopped-building-webhook-retry/https://blog.axme.ai/stopped-building-webhook-retry/Exponential backoff, jitter, dead letter queues, idempotency keys, HMAC verification - all to deliver one message reliably. There are better options now.Sun, 29 Mar 2026 00:00:00 GMTcloudwebhookspythonbackendarchitecturehttps://blog.axme.ai/agent-crashed-at-step-47/https://blog.axme.ai/agent-crashed-at-step-47/AI agents do multi-step work. They crash. The state is gone. Here's how to make agent operations durable without writing checkpoint code.Thu, 26 Mar 2026 00:00:00 GMTcloudaipythonagentsdurabilityhttps://blog.axme.ai/agent-shouldnt-block-on-approval/https://blog.axme.ai/agent-shouldnt-block-on-approval/AI agents get stuck waiting for humans. There's a better pattern than blocking - async approval with reminders, escalation, and timeout.Thu, 26 Mar 2026 00:00:00 GMTcloudaipythonagentshumanintheloophttps://blog.axme.ai/what-temporal-cant-do/https://blog.axme.ai/what-temporal-cant-do/Temporal gives you durable execution. But adding a human approval step mid-workflow requires building a signal handler, notification service, and UI. There's a simpler way.Thu, 26 Mar 2026 00:00:00 GMTcloudtemporalpythonworkflowdurableexecution