Aigon feature sets are more than a grouping label. In autonomous mode, a set becomes the unit of orchestration: Aigon walks the member features in dependency order, runs each feature through the existing autonomous loop, and carries the run forward until the set is complete or paused.

If you want a concrete mental model, the Brewboard tutorial shows the single-feature autonomous flow in a small sandbox repo. This page explains how that same idea scales from one feature to a whole set.

What a feature set is

A feature set is a shared set: <slug> frontmatter value across multiple feature specs. The set is derived from the specs themselves; there is no separate hidden database for membership.

The important part is that the set gives Aigon a stable cluster of related work:

• related features can be grouped together on the board
• the dashboard can show aggregate progress for the cluster
• autonomous execution can treat the set as a sequenced queue

How autonomous mode uses it

When you start a set autonomously, Aigon does not invent a new implementation loop. It reuses the existing per-feature autonomous machinery and runs it member by member.

The set conductor:

• resolves the set members from set: <slug> frontmatter
• topologically sorts them using the existing dependency graph
• starts each member with the normal feature-autonomous-start flow
• waits for each member to finish before moving on, unless the set is configured to behave otherwise
• persists set-level run state so it can resume after interruption or failure

That means the set behavior stays aligned with the single-feature workflow. You get orchestration across the set without duplicating the logic that already governs a feature run.

Why this matters

The autonomous set flow is useful when one feature is really a cluster of smaller tasks that should land in order.

It helps when you want to:

• kick off a whole project slice once and come back later
• avoid manually restarting the next feature after each close
• keep dependency order explicit instead of relying on memory
• resume a paused set without rebuilding the entire plan

In practice, feature sets let Aigon operate at a higher level of abstraction while still keeping each feature’s own lifecycle intact.

Relationship to the dashboard

The dashboard shows feature sets as cards with progress and action buttons. That makes the current set state visible while the conductor is running.

The dashboard view is read-only. The actual set behavior comes from the workflow engine and the set conductor, not from frontend heuristics.

Brewboard as the example

Use Brewboard to demonstrate the idea in a low-risk repo:

• a single feature run shows the base autonomous lifecycle
• a feature set lets you chain several related Brewboard specs in dependency order
• the dashboard then shows the set card while the conductor advances through the members

That makes Brewboard a good teaching repo because it keeps the moving parts small while still showing the real autonomous control flow.

See also

• Tutorial: Brewboard Feature Set
• Tutorial: Try Aigon with Brewboard
• Dashboard
• Feature Sets reference
• Feature set research