Operator papers for the era of man and machine.

Aweb Study 001: a public operator paper on serious AI work, receipts, constrained agents, and the environment around human-machine execution.

A plain technical note on the agent loop: context, planning, tools, observation, evaluation, and human boundaries.

The operating brief I want before a system touches tools: mission, authority, constraints, evidence, and stop conditions.

Why a confident answer is not enough, and how Aweb uses traces, corrections, and artifacts to keep work inspectable.

A visual Aweb architecture study: operator interface, policy, orchestration, agents, tools, runtime, and receipts.

How tool connectors let agents discover, authorize, execute, and observe actions without pretending every integration is magic.

Aweb's view of APIs as context-aware, policy-driven, composable, observable execution surfaces for agentic systems.

Two public Aweb board sets covering agent operating systems, loops, MCP, receipts, orchestration, memory, review, security, collaboration, and mission lifecycle.

How Aweb thinks about provider capabilities, routing, policies, cost, latency, evidence, and tool selection under mission constraints.

The technical shape of inspectable machine work: mission IDs, tool traces, artifacts, evaluations, corrections, and review.

Why agent memory has to be paired with evaluation, correction, permission, and human review before it becomes useful.

A public map of Aweb, GEX, Veritas, Nina, and Leony as one work graph with separate domains and claim boundaries.