Operating manual
Tag: Agent Operations
A focused reading path for Agent Operations: related field notes, evidence trails, and operating questions from the archive.
- more-compute-wont-fix-local-computer-use-agents.md
Four Compute Bets for Computer-Use Agents
For computer-use agents, context, steps, decomposition, and parallel plans are different bets. Here is how to test which one earns its cost.
open artifact → - system-prompts-are-product-policy.md
System Prompts Are Product Policy
When a prompt clause changes consequential AI behavior, it needs a release record—not an informal text edit.
open artifact → - more-coding-agents-can-make-the-work-worse.md
More Coding Agents Can Make the Work Worse
When do multiple coding agents help? Use this five-gate test to decide whether parallel work justifies its coordination tax.
open artifact → - coding-agent-recovery-is-a-budget-decision.md
Coding-Agent Recovery Is a Budget Decision
A failed coding-agent run creates evidence. Use a Recovery Budget Router to decide whether to repair, replan, escalate, or stop.
open artifact → - self-hosting-ai-is-a-control-decision.md
Self-Hosting AI Is a Control Decision
A five-part Control Boundary Test for deciding whether an AI workload belongs on self-hosted, managed, or hybrid infrastructure.
open artifact → - vertical-ai-needs-receipts-not-better-chat.md
Vertical AI Needs Receipts, Not Better Chat
A five-part scorecard for testing whether vertical AI can use real tools, preserve execution context, produce inspectable artifacts, and support reruns.
open artifact →