Operating manual
Tag: Agent Operations
A focused reading path for Agent Operations: related field notes, evidence trails, and operating questions from the archive.
- ai-memory-needs-an-eviction-policy.md
AI Memory Needs an Eviction Policy
A memory budget sets capacity. A memory policy decides what earns retention—and whether the rule actually helps the workload.
open artifact → - agent-work-needs-chain-of-custody-not-just-bigger-models.md
Agent Work Needs Chain-of-Custody, Not Just Bigger Models
A four-question acceptance test for trusting agent work: who authorized it, what state crossed the handoff, what evidence proves completion, and what recovery path exists.
open artifact → - agent-work-needs-chain-of-custody-not-just-capability.md
Agent Work Needs Chain-of-Custody, Not Just Capability
A four-question acceptance test for trusting agent work: who authorized it, what state crossed the handoff, what evidence proves completion, and what recovery path exists.
open artifact → - the-authority-envelope-every-agent-handoff-should-carry.md
The Authority Envelope Every Agent Handoff Should Carry
Agent handoffs do not become trustworthy because two systems can talk. They become trustworthy when every handoff carries authority, state, evidence, risk, and the next allowed action.
open artifact → - the-next-agent-bottleneck-is-operational-control-not-model-capability.md
The next agent bottleneck is operational control, not model capability
Frontier models keep improving. The agent bottleneck is moving to state, permissions, evals, recovery, and proof that work actually shipped.
open artifact → - green-status-is-not-shipped-work.md
Green Status Is Not Shipped Work
Agent workflows need deliverable-level acceptance tests because internal completion state is not the same thing as user-facing proof.
open artifact →