Overview
The agentic workflow is how a change at ttoss goes from a signal to an effect observed in production. Three ideas shape it: it is a state machine over a task graph, one engine runs it for every project while each project extends it, and it is built on Get Things Flowing (GTF) and Flow.
Principles
The agentic workflow rests on three pillars. The product development principles say what the work is worth and why flow beats utilization. The agentic development principles say how AI agents behave: context resets, output is probabilistic, and a proxy optimized hard enough stops tracking the goal. The workflow's own principles combine the two into a design. Every rule of the Delivery Process, the parameters and the routines applies one of them.
Delivery Process
How work in a software project goes from a signal to a closed task, once, for every project. It is a state machine the rule ids below are GTF's.
Project Parameters
The delivery process is the engine, the same for every project; these are its extension points, and a project's answers to them are how it delivers. The tick reads them from the consuming repo, first match wins: the /flow argument (project id only), .agents/flow.md, .claude/flow.md, or a ## Flow parameters section of AGENTS.md / CLAUDE.md. /flow-setup defines them with the user. They bind every session working in the repo, not only the agents'.
Routines
A routine is a run that a schedule or an event starts, not a person. Each one runs like a tick: it starts fresh, reads the state from Flow, acts, writes the result back and stops, so a run that fails loses nothing and the next one picks up from the graph (GTF-U4). The runtime fires them, as a Claude Code routine or a cron or CI job that runs the agent headless, and the project lists them in its Routines parameter.