How to Configure Flags in Cubyts?

Introduction

This guide explains how to configure Flags in the Cubyts platform to control governance behavior, automation, and traceability. Flag configuration allows teams to fine-tune how risks are detected, how actions are automated, and how governance signals are captured for audits—ensuring that AI-driven insights align with organizational standards and intent.

Prerequisites

• Integrated planning and code repositories

• Understanding of process, feature, and code governance goals

Step-by-Step Guide

Step 1: Open the Flag Configuration Screen

1. Navigate to Flag Configuration in your Cubyts workspace.

2. This screen is used to manage:

   • Flag visibility

   • Flag behavior

   • Automation and audit settings

Any configuration changes take effect after the next sync cycle, ensuring predictable and controlled updates.

Step 2: Understand Flag Categories and Indicators

• Flags are grouped into three categories:

  • Process flags

  • Feature flags

  • Code flags
    
    

• Additional indicators show whether a flag is:

  • Audit-relevant

  • Auto-resolvable

  • AI-powered

This structure helps you quickly locate and manage the right governance checks.

Step 3: Enable or Disable Flags

• Each flag has a toggle switch:

  • On – Enables evaluation and surfacing of insights

  • Off – Pauses the flag without deleting its configuration

• This allows teams to:

  • Roll out governance incrementally

  • Temporarily disable checks without losing context

Step 4: Create Custom Code Flags

• Select Create Custom Flag from the top-right corner of the configuration screen.

• Custom flags are supported only for code flags.

To create a custom code flag:

1. Provide a flag name and description.

2. Define conditions for the AI engine to evaluate pull requests or branches.

3. Optionally simulate outcomes using an existing open pull request.

Example use cases:

• Performance checks

• Logic checks

• Organization-specific security/compliance checks

• Organization-specific coding standards

This enables custom governance without impacting planning or process workflows.

Step 5: Configure Limit Settings

Limit settings control how and when a flag surfaces.

For example (for one of the Cubyts flags):

• Assign weightages to contributing parameters such as:

  • Team member overload

  • Requirement and design quality

  • Scope changes during sprint

  • Build planning finesse

• The cumulative weightage always totals 100%.

• Define a threshold percentage that determines when the risk is high enough for the flag to trigger.

• Use a time-based control to specify how far into a sprint the flag should activate.

These controls help tune sensitivity and reduce noise.

Step 6: Configure Auto-Resolution Behavior

Auto-resolution settings define how flags can be resolved automatically.

For example (for one of the Cubyts flags)

• For a sprint overrun prediction flag, Cubyts can:

  • Automatically move a work item to another sprint

  • Allow configuration of the target sprint

Project-specific mappings determine how new work items are created when needed, reducing manual follow-ups and enabling proactive correction.

Step 7: Configure Audit Settings

• Enable Audit Settings to mark a flag as audit-relevant.

• When enabled:

  • Insights are captured in audit trails

  • Signals flow into governance and audit dashboards

This supports both internal governance reviews and external compliance audits.

Step 8: Define Benchmarks for Build Tasks

Benchmarks define what good looks like for build planning and execution.

1. Configure benchmarks for build types such as:

   • UI builds

   • API builds

   • Backend builds

   • Database builds

2. Each benchmark links to a reference artifact (for example, a well-planned Jira or Azure Boards work item).

3. Multiple benchmarks can be added per build type.

Process and feature flags use these benchmarks to detect deviations in planning and execution.

Step 9: Define Benchmarks for Requirement Quality

Requirement benchmarks establish standards for:

• Functional requirements

• Technical requirements

• Deep technical requirements (frontend, backend, database)
  
  

Each benchmark references an artifact that demonstrates:

• Expected structure

• Required detail

• Documentation quality

This ensures requirement quality is evaluated against your organization’s standards, not generic rules.

Step 10: Configure Code Flag Behavior

Code flag configuration controls how pull requests and branches are evaluated.

You can configure:

• Whether flags apply to:

  • Pull requests

  • Branches

  • Both

• Time-based thresholds (for example, aging PRs or long-running branches)

• Optional visibility of flags directly in:

  • GitHub

  • GitLab

  • Bitbucket

Auto-resolution rules can also be defined to resolve flags once conditions are met.

The Common Flag Configuration Model

Every flag in Cubyts—process, feature, or code—is built on the same core pillars:

1. Limit settings – When and how a flag surfaces
   
   

2. Auto-resolution settings – How actions are automated

3. Audit settings – How signals are captured for traceability

Benchmarks extend this model by defining quality baselines, while code flag configuration applies the same principles to repositories and pull requests.

Conclusion

Flag configuration in Cubyts enables teams to scale governance intelligently—without increasing operational overhead. By combining configurable limits, automation, audits, and benchmarks, teams can align AI-driven signals with real governance intent, ensuring consistent, explainable, and traceable oversight across planning, delivery, and code.

Video link: https://www.loom.com/share/6496f33b0e034e45925911e4ad421000