Skip to content

SASE Blog

The SASE Blog is the publishing surface for essays about Structured Agentic Software Engineering (SASE): durable work units, orchestration, provider-independent workflows, review state, and the coordination layer around coding agents.

Start Here

Start with Hello, SASE: Your First 15 Minutes if you want the practical path: install, provider readiness, a safe first run, and the first visible agent record.

The SASE Blog Series also begins with [00] The Missing Operating Layer for Coding Agents, the launch essay on why coding-agent work needs durable prompts, plans, state, review, dependencies, retries, and handoff.

More posts in the series are in the repository as drafts and will be published over time. The generated archive below lists only the entries included in the public site.

From Reading To Practice

  • Start with ACE, the Agentic ChangeSpec Explorer, to learn the terminal interface for daily agent work.
  • Read the SDD flow to see how Spec-Driven Development turns plans, epics, and phase beads into executable work.
  • Open the repository for source, issues, and implementation details.
  • 6 min read

[01] Hello, SASE — Your First 15 Minutes Orchestrating Coding Agents

SASE (pronounced "sassy" — yes, really) is a coordination layer that sits above coding-agent CLIs like Claude Code, Codex, or Antigravity CLI (agy). This post is the practical on-ramp: by the end you'll have installed sase, checked that a provider CLI is ready, launched a safe read-only agent run, found the resulting agent record, and picked up the vocabulary you'll keep bumping into in the rest of the docs. Plan on roughly fifteen minutes at a terminal, plus however long your favorite model takes to think.

  • 17 min read

[00] The Missing Operating Layer for Coding Agents

SASE is not a better model. SASE is the layer I wanted after realizing that the hard part of running coding agents is not always "can the model write the patch?" Sometimes the hard part is "where did the patch go?", "what was it trying to do?", "who is waiting on it?", "why did it start six follow-up agents while I was brushing my teeth?", and "can I please see the diff before the robot commits crimes against just check?"

That layer needs reusable prompts, durable plans, dependency-aware work items, review records, background automation, notifications, and a control surface that lets humans steer without becoming a full-time air-traffic controller.

Borrowing the name from the research paper discussed later, SASE calls that layer Structured Agentic Software Engineering. This post is the map of the fundamentals: XPrompts, SDD, Beads, ACE, AXE, plugins, and why SASE wraps coding-agent CLIs instead of raw model APIs.