Published April 25, 2026 | Version 1.0-preprint
Working paper Open

Operating Model for Agentic Teams: Standup, Sprint Planning, and Context Discipline When the Dev Team Is Non-Human

Authors/Creators

  • 1. Non Sequitur Publishing

Description

Agile Scrum gave software development a teachable operating model — but the model presumes biological team members with continuous memory, ambient context propagation through co-presence, and conversation as the carrier of decisions. When the Dev Team is non-human, all three assumptions fail at once. Modern agentic platforms supply capability — language models, agent harnesses, tool integrations — but they do not specify how a team-sized unit operates around them.

This paper specifies that operating model. It is the day-to-day form of the Skipjack Protocol applied to a team-sized unit of work, occupying the Application layer of the four-tier governance flow the Mission-Ready AI decalogy describes (Framework HGC³AE² → Protocol Skipjack → Harness → Application). The model is teachable: it has named ceremonies, named artifacts, named gate criteria, and a defined cadence.

Four claims structure the paper. Ceremony cadence is structural, not optional — the standup analogue, the sprint-planning analogue, and the retrospective analogue exist because operator comprehension decays between sessions and must be re-established by routine. The orchestrator is the admission-gate arbiter — no task enters the Agentic Execution Layer without verification that it carries a Governance-Layer-approved decomposition envelope. Stories and tasks layer over HG+C³ and AE² as first-class governance artifacts — a story is a curation envelope plus a context contract, and a task is an admission-gated, evidence-producing decomposition. Context-boundary maintenance is implemented as a multi- account split — cognitive modes are isolated into distinct agent contexts, with re-unification at named decision points and drift detection as a continuous routine.

Paper Six is the sixth paper in the Mission-Ready AI decalogy and the first of three Application-layer depth papers. The operating model the paper describes is the model under which the decalogy itself was written; the working portfolio is its existence proof.

Rights envelope: Citation permitted with full attribution. No reproduction, redistribution, or derivative works without written permission. AI/ML training use disallowed. See the citation policy at https://nonsequitur.tech/pubs/citation-policy/ for the full rights envelope.

Canonical site URL: https://nonsequitur.tech/white-papers/operating-model-agentic-teams/

Public archive: yks-pubs/papers/operating-model-agentic-teams-v1-preprint.pdf

Files

operating-model-agentic-teams-v1-preprint.pdf

Files (2.5 MB)

Name Size Download all
md5:62518482afd9f271abdb7be56f905f1a
2.5 MB Preview Download

Additional details