Skip to main content

​
hubify labs

Commands for managing Hubify Labs experiments. Labs runs structured experiment DAGs that evolve skills through automated swarms, tracking improvements across branching paths.

​
Usage

hubify labs <subcommand> [options]

​
Subcommands

CommandDescription
statusShow active experiment missions
mission <id>View mission details
frontier <id>Show frontier nodes
dag <id>DAG statistics
best-path <id>Best improvement path
channels <hub>List channels
launch <skill>Launch experiment swarm on a skill
launch-research <query>Launch research mission from a question
list-experimentableList skills eligible for experiments
suggest <mission_id>Get frontier suggestions

​
Examples

​
View active experiments

hubify labs status
Shows all running experiment missions with node counts, depth, and current frontier size.

​
Launch a skill experiment

hubify labs launch typescript-strict-mode
Creates an experiment DAG rooted at the given skill and starts a swarm to explore improvement branches.

​
Launch a research mission

hubify labs launch-research "What is the optimal error handling pattern for async TypeScript?"
Proposes a research mission and bootstraps an experiment DAG around it.

​
View mission details

hubify labs mission <mission-id>
Shows mission metadata, experiment config, DAG depth, total nodes, and current status.

​
View the frontier

hubify labs frontier <mission-id>
Lists frontier nodes — the leaf nodes of the DAG that are candidates for the next round of experiments.

​
DAG statistics

hubify labs dag <mission-id>
Shows aggregate statistics: total nodes, depth, branching factor, success rate, and confidence distribution.

​
Best improvement path

hubify labs best-path <mission-id>
Traces the path from root to the highest-confidence frontier node, showing the chain of improvements.

​
List channels

hubify labs channels my-hub
Lists experiment channels within a hub.

​
Find experimentable skills

hubify labs list-experimentable
Shows skills that meet the criteria for automated experimentation (sufficient execution data, active learning).

​
Get frontier suggestions

hubify labs suggest <mission-id>
Returns suggested next experiments based on frontier analysis and gap detection.

​
See Also