What is the CC/CD framework?

Continuous Calibration/Continuous Development (CC/CD) is a methodology focused on the iterative testing, monitoring, and refinement of AI behaviors to ensure safety and quality as autonomy increases.

What is an Agency Ladder for AI?

An Agency Ladder is a framework for mapping how an AI feature transitions from human-controlled suggestions (Low Agency) to autonomous actions (High Agency) based on proven reliability.

When should I promote an AI feature to the next level of autonomy?

Promotion occurs when quality metrics are stable, error patterns are understood, user correction rates are decreasing, and operational monitoring is fully in place.

Is it okay for a feature to stay at V1 indefinitely?

Yes. If the consequences of an AI error are too high or if human judgment is critical to the task, maintaining a high-control, low-agency (V1) state is a valid and safe product decision.

Agency Ladder

Name: Agency Ladder
Author: breethomas

bybreethomas

•

Productivity & Workflow

Standardizes the progression of AI feature autonomy from human-in-the-loop to autonomous action.

About

Agency Ladder is a specialized product management skill designed to help teams systematically plan, review, and promote AI features through increasing levels of autonomy. Grounded in the CC/CD (Continuous Calibration/Continuous Development) framework, it provides a structured path for moving AI capabilities from low-agency suggestions (V1) to high-agency autonomous actions (V3). By enforcing clear promotion criteria and documenting control handoffs, this skill ensures that AI earns its autonomy through proven reliability rather than being granted it by default, reducing risk and improving stakeholder alignment.

Key Features

Structured V1-V2-V3 autonomy roadmap planning
10 GitHub stars
Existing feature gap analysis and risk assessment
Automated promotion criteria checklists for AI reliability
Interactive command-line workflows for rapid planning
Presentation-ready stakeholder alignment artifacts

Use Cases

Auditing existing AI features to determine if they are ready for increased autonomy
Designing the phased rollout of a new LLM-powered feature
Generating governance documentation for AI safety and human-in-the-loop controls

About

Key Features

Structured V1-V2-V3 autonomy roadmap planning
10 GitHub stars
Existing feature gap analysis and risk assessment
Automated promotion criteria checklists for AI reliability
Interactive command-line workflows for rapid planning
Presentation-ready stakeholder alignment artifacts

Use Cases

Auditing existing AI features to determine if they are ready for increased autonomy
Designing the phased rollout of a new LLM-powered feature
Generating governance documentation for AI safety and human-in-the-loop controls