--- title: "#agents articles | Hashnode" description: "Explore 1,100 articles tagged #agents on Hashnode. Tutorials, guides, and how-tos from the community." url: "https://hashnode.com/tag/agents" --- Tag # #agents 1,100 posts · 16 followers [MemorySync](https://hashnode.com/@memorysync) · [2h ago](https://hashnode.com/posts/langgraph-durability-sync-no-attribute-put-checkpoint-fut/6af888c31b05f2a033c918be) ### [LangGraph durability='sync' without a checkpointer: fixing 'SyncPregelLoop' object has no attribute '\_put\_checkpoint\_fut'](https://memorysync.hashnode.dev/langgraph-durability-sync-no-attribute-put-checkpoint-fut?utm_source=hashnode&utm_medium=feed) Pass durability='sync' to a LangGraph graph compiled without a checkpointer and the run crashes after the first step. Here's why, and three small workarounds until a fix ships. 12 min read · [memorysync.hashnode.dev](https://memorysync.hashnode.dev/langgraph-durability-sync-no-attribute-put-checkpoint-fut?utm_source=hashnode&utm_medium=feed) [MemorySync](https://hashnode.com/@memorysync) · [2h ago](https://hashnode.com/posts/n8n-zep-memory-resource-you-are-requesting-could-not-be-found/6af887322c1b2855faeedf8b) ### [n8n Zep memory: fixing 'The resource you are requesting could not be found' (and replacing the deprecated node)](https://memorysync.hashnode.dev/n8n-zep-memory-resource-you-are-requesting-could-not-be-found?utm_source=hashnode&utm_medium=feed) The Zep credential test calls an endpoint Zep retired, so it fails even with a good key. Saving anyway can look like it works while history quietly stays empty. Here's what's happening and how to move the agent to a maintained memory node. 13 min read · [memorysync.hashnode.dev](https://memorysync.hashnode.dev/n8n-zep-memory-resource-you-are-requesting-could-not-be-found?utm_source=hashnode&utm_medium=feed) [Hao | PMNote](https://hashnode.com/@pmnote) · [1d ago](https://hashnode.com/posts/the-ai-shift-through-the-eyes-of-an-online-game-pm/6af73fe2c5607a9ba23194f4) ### [The AI shift through the eyes of an online game PM](https://pmnote.hashnode.dev/the-ai-shift-through-the-eyes-of-an-online-game-pm?utm_source=hashnode&utm_medium=feed) Person-day estimates, linear pipelines, daily stand-ups, long documents, headcount-heavy QA, and gut-feel judgment: six old agreements are being broken by AI, because the premise that “people are the 7 min read · [pmnote.hashnode.dev](https://pmnote.hashnode.dev/the-ai-shift-through-the-eyes-of-an-online-game-pm?utm_source=hashnode&utm_medium=feed) [MemorySync](https://hashnode.com/@memorysync) · [1d ago](https://hashnode.com/posts/crewai-memory-checkpoint-unable-to-serialize-unknown-type-modelmetaclass/6af7368de0761f91261ee465) ### [CrewAI memory + checkpointing: fixing 'Unable to serialize unknown type: ModelMetaclass' (no checkpoint written)](https://memorysync.hashnode.dev/crewai-memory-checkpoint-unable-to-serialize-unknown-type-modelmetaclass?utm_source=hashnode&utm_medium=feed) Turn on Memory and checkpointing together and every task\_completed checkpoint silently fails to write. Here's why, how to catch it, and a small patch until the fix ships. 11 min read · [memorysync.hashnode.dev](https://memorysync.hashnode.dev/crewai-memory-checkpoint-unable-to-serialize-unknown-type-modelmetaclass?utm_source=hashnode&utm_medium=feed) [Decillion team](https://hashnode.com/@decillionai) · [1d ago](https://hashnode.com/posts/a-reader-s-idea-for-a-second-spend-limit-on-ai-agents/6af729cc5d26b28c671be2c6) ### [A reader's idea for a second spend limit on AI agents](https://decillionai.hashnode.dev/a-reader-s-idea-for-a-second-spend-limit-on-ai-agents?utm_source=hashnode&utm_medium=feed) Budgets cap how much. Rate limits would cap how fast. Here's our thinking, and what we'd like to hear. 3 min read · [decillionai.hashnode.dev](https://decillionai.hashnode.dev/a-reader-s-idea-for-a-second-spend-limit-on-ai-agents?utm_source=hashnode&utm_medium=feed) [Abdus Salam](https://hashnode.com/@aheadkit) · [1d ago](https://hashnode.com/posts/a2ui-1-0-counting-what-the-21-token-rate-card-really-saves/6af6728d8b771824d3f489f6) ### [A2UI 1.0: Counting What the 21-Token Rate Card Really Saves](https://aheadkit.hashnode.dev/a2ui-1-0-counting-what-the-21-token-rate-card-really-saves?utm_source=hashnode&utm_medium=feed) A2UI 1.0 shipped with a striking number attached to it. A currency agent built on both versions of the protocol shows the same rate card costing the model 223 tokens to write on A2UI 0.9 and 21 tokens 8 min read · [aheadkit.hashnode.dev](https://aheadkit.hashnode.dev/a2ui-1-0-counting-what-the-21-token-rate-card-really-saves?utm_source=hashnode&utm_medium=feed) [Irfan Gulzar](https://hashnode.com/@irfangulzarx) · [1d ago](https://hashnode.com/posts/ghost-agents-82-of-companies-are-running-ai-agents-they-never-approved/6af6d2e5e9abe111937f8109) ### [Ghost Agents: 82% of Companies Are Running AI Agents They Never Approved](https://irfangulzarx.hashnode.dev/ghost-agents-82-of-companies-are-running-ai-agents-they-never-approved?utm_source=hashnode&utm_medium=feed) Most companies can tell you exactly how many employees they have. Ask how many AI agents are running inside their systems right now — reading email, touching the CRM, moving data between tools — and t 8 min read · [irfangulzarx.hashnode.dev](https://irfangulzarx.hashnode.dev/ghost-agents-82-of-companies-are-running-ai-agents-they-never-approved?utm_source=hashnode&utm_medium=feed) [Vedar](https://hashnode.com/@vedarai) · [1d ago](https://hashnode.com/posts/two-layers-one-task-how-mcp-and-a2a-actually-fit-together/6af6c79b69ffc4562866f3da) ### [Two layers, one Task: how MCP and A2A actually fit together](https://vedarai.hashnode.dev/two-layers-one-task-how-mcp-and-a2a-actually-fit-together?utm_source=hashnode&utm_medium=feed) Serious AI agents now sit on two protocols that do different jobs. The Model Context Protocol (MCP) is how an agent reaches tools and data. The Agent2Agent protocol (A2A) is how one independent agent 6 min read · [vedarai.hashnode.dev](https://vedarai.hashnode.dev/two-layers-one-task-how-mcp-and-a2a-actually-fit-together?utm_source=hashnode&utm_medium=feed) [Shivam Kumar](https://hashnode.com/@shivaysinghrajput) · [1d ago](https://hashnode.com/posts/the-agent-reliability-stack-is-the-next-big-build-here-s-the-opening/6af694a142a9fc697fb55673) ### [The Agent Reliability Stack Is the Next Big Build — Here's the Opening](https://shivaysinghrajput.hashnode.dev/the-agent-reliability-stack-is-the-next-big-build-here-s-the-opening?utm_source=hashnode&utm_medium=feed) The Agent Reliability Stack Is the Next Big Build — Here's the Opening Shivam Kumar | VisionQuantech | November 13, 2026 Every gold rush has its shovel sellers. The AI agent gold rush already has its 5 min read · [shivaysinghrajput.hashnode.dev](https://shivaysinghrajput.hashnode.dev/the-agent-reliability-stack-is-the-next-big-build-here-s-the-opening?utm_source=hashnode&utm_medium=feed) [Veronica Stewart](https://hashnode.com/@veronicstewart) · [2d ago](https://hashnode.com/posts/stop-approving-one-by-one-batch-it/6af602a77e155c8935ef3cec) ### [Stop approving one by one. Batch it.](https://veronicstewart.hashnode.dev/stop-approving-one-by-one-batch-it?utm_source=hashnode&utm_medium=feed) Originally published on DEV: https://dev.to/veronicstewart/stop-approving-one-by-one-batch-it-412d Per-action approval feels safe. At seven actions it already isn't. I run approval-gated agent work in 3 min read · [veronicstewart.hashnode.dev](https://veronicstewart.hashnode.dev/stop-approving-one-by-one-batch-it?utm_source=hashnode&utm_medium=feed) [2](https://hashnode.com/posts/stop-approving-one-by-one-batch-it/6af602a77e155c8935ef3cec)