https://triadive.com A working field manual for the human / agent / robot triad. Concepts, prompts, plugins, and lessons, written at the level of principles. en-us Mon, 31 Aug 2026 08:04:45 +0000 https://triadive.com/dispatches/2026-08-29-issue-5-the-ceo-agent-bottleneck/ https://triadive.com/dispatches/2026-08-29-issue-5-the-ceo-agent-bottleneck/ Adding more desks does not make a multi-agent system faster if every desk still reports to the same overloaded coordinator. The CEO-agent bottleneck is the patt Sat, 29 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/field-notes/work-desks-and-specialty-agents/ https://triadive.com/articles/field-notes/work-desks-and-specialty-agents/ Sessions, jobs, work desks, and specialty agents are four names for the same primitive — and knowing which one you mean changes how you design, debug, and gover Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/from-prompts-to-agents-the-gap-youre-probably-skipping/ https://triadive.com/articles/concepts/from-prompts-to-agents-the-gap-youre-probably-skipping/ Most people who use LLMs daily are running one-shot or multi-turn prompts — not agents. The difference matters more than it sounds. Here is what actually change Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/from-prompts-to-desks/ https://triadive.com/articles/concepts/from-prompts-to-desks/ The first time an agent system is set up, it is a single agent in a single session doing everything. That works until it does not. This is the beginner bridge f Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/what-is-an-llm/ https://triadive.com/articles/concepts/what-is-an-llm/ Large language models are next-token predictors trained on a wide corpus of human-written text. This piece explains what the model actually does, how it is trai Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/workflows/setting-up-your-first-three-desks/ https://triadive.com/articles/workflows/setting-up-your-first-three-desks/ Three desks cover most small operations: a research desk, an editorial or production desk, and an operations desk. This workflow walks through the setup step by Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/workflows/setting-up-your-first-three-work-desks/ https://triadive.com/articles/workflows/setting-up-your-first-three-work-desks/ The most common mistake in early agent architecture is building one agent that does everything, then watching it slowly lose coherence under the weight of compe Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/lessons/desk-architecture-failure-modes-what-breaks-first-and-how-to-fix-it/ https://triadive.com/articles/lessons/desk-architecture-failure-modes-what-breaks-first-and-how-to-fix-it/ After running desk-based architectures across dozens of agent systems, the same failure patterns show up in the same order: context bleeds, memory饱和, handoff dr Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/lessons/desk-architecture-failure-modes/ https://triadive.com/articles/lessons/desk-architecture-failure-modes/ Adding desks does not automatically make a system better. It can make it worse — slower, more confused, and harder to debug. This lesson catalogs the six failur Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/glossary/job/ https://triadive.com/glossary/job/ A specific assignment given to a work desk: a unit of work with a name, inputs, a deadline, and an expected output. The job is to the desk what a task is to a s Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/glossary/specialty-agent/ https://triadive.com/glossary/specialty-agent/ A work desk tuned for a narrow function: a specific role, a specific toolset, and a specific scope of work. A specialty agent is a work desk with a defined spec Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/glossary/work-desk/ https://triadive.com/glossary/work-desk/ A durable operating context where a named line of work lives: a goal, a role, a memory surface, selected tools, a job queue, and rules for escalation. In a heal Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/glossary/workboard/ https://triadive.com/glossary/workboard/ A card-based state-tracking system for agent operations. A workboard card holds status, assignee, priority, and an audit trail — the system of record for what i Wed, 26 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/dispatches/2026-08-22-issue-4-the-vocabulary-tax-when-primitives-go-unnamed/ https://triadive.com/dispatches/2026-08-22-issue-4-the-vocabulary-tax-when-primitives-go-unnamed/ When the components of an agent system each use a different word for the same thing, the operator pays the cost in slow debugging, missed audits, and prompts th Sat, 22 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/dispatches/2026-08-15-issue-3-why-bounded-context-wins/ https://triadive.com/dispatches/2026-08-15-issue-3-why-bounded-context-wins/ The most expensive failure mode in long-running agents is not bad planning, bad tools, or bad models. It is unbounded context — the slow accumulation that makes Sat, 15 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/field-notes/field-note-2026-08-09-when-prompt-cache-saved-the-morning/ https://triadive.com/articles/field-notes/field-note-2026-08-09-when-prompt-cache-saved-the-morning/ A small prompt restructure cut our morning cron latency from 4s to 1.1s and dropped the daily cost by 14%. Nothing else changed. A field note on what we learned Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/agents-and-robotics/ https://triadive.com/articles/concepts/agents-and-robotics/ A software agent that sends a wrong email is reversible. A robot that drives into a wall is not. This piece covers what changes when the agent's actions touch t Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/context-management-and-system-prompts/ https://triadive.com/articles/concepts/context-management-and-system-prompts/ A system prompt is not a single file. It is assembled at runtime from a project file, an agent default, a tool-result injection, and the conversation history. T Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/evaluation-safety-and-governance/ https://triadive.com/articles/concepts/evaluation-safety-and-governance/ Agents fail in ways classical software does not. Eval has to catch open-ended, multi-step, tool-mediated behavior. Safety has to bound runaway loops, prompt inj Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/first-steps-sessions-and-memory/ https://triadive.com/articles/concepts/first-steps-sessions-and-memory/ A session is a single conversation. Memory is what survives between sessions. Together they form the agent's lifecycle: each session starts cold, runs hot, and Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/from-prompts-to-agents/ https://triadive.com/articles/concepts/from-prompts-to-agents/ A prompt is a one-shot instruction. An agent is a system that runs prompts over time. The move from one to the other is not a single jump — it is a series of ad Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/future-of-agent-architectures/ https://triadive.com/articles/concepts/future-of-agent-architectures/ Today's agent architecture is mostly language model plus memory plus tools plus loop. It works and will not be the final form. This piece is a forward-looking m Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/getting-started-with-workspaces/ https://triadive.com/articles/concepts/getting-started-with-workspaces/ A workspace is the file system plus memory plus agent identity that an agent reads when it wakes up. The shape of that workspace determines what the agent can d Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/how-does-an-agent-understand/ https://triadive.com/articles/concepts/how-does-an-agent-understand/ An agent 'understands' by routing a query across three places: the current context window, the memory layer, and the model's pattern-matching weights. None of t Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/practical-applications/ https://triadive.com/articles/concepts/practical-applications/ Three generations of agent deployments — personal, team, business — differ not in capability but in blast radius, governance, and cost model. This piece is a pr Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/prompting-patterns-and-anti-patterns/ https://triadive.com/articles/concepts/prompting-patterns-and-anti-patterns/ Most agent failures are prompting failures in disguise. The agent did not follow instructions, did not stay on task, did not produce the right output — not beca Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/sandbox-and-terminal-safety/ https://triadive.com/articles/concepts/sandbox-and-terminal-safety/ An agent that can run shell commands is an agent that can destroy the file system, leak secrets, or call external APIs in ways the operator did not authorize. T Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/what-is-ai/ https://triadive.com/articles/concepts/what-is-ai/ AI is a marketing label that covers several distinct technical generations — rules, classical ML, deep learning, LLMs, agents. The words mean different things a Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/concepts/workspace-organization-and-project-hygiene/ https://triadive.com/articles/concepts/workspace-organization-and-project-hygiene/ A well-run agent workspace is a small set of explicit files: AGENTS.md, SOUL.md, TOOLS.md, MEMORY.md, and the daily notes. Each file has a job; each file has a Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial https://triadive.com/articles/workflows/an-agent-eval-workflow/ https://triadive.com/articles/workflows/an-agent-eval-workflow/ A practical recipe for evaluating an agent system before it ships. Four layers to test, a 30-line harness in prose, what to do when an eval fails, and a minimum Sun, 09 Aug 2026 00:00:00 +0000 Triadive Editorial