The tool riskratchet has been introduced for Python, designed to protect codebases from quality degradation during active use of AI agents. The tool utilizes a "ratchet" mechanism, allowing users to set baseline quality levels and block changes that lead to increased technical debt.
What Happened
A Python tool called riskratchet has been developed, which calculates a combined risk score for each function. The assessment is based on several metrics: coverage gap, structural complexity (cyclomatic complexity), branch gap, churn (frequency of changes), public surface area, and overall code sprawl. The system allows setting a current baseline and automatically blocks commits or CI builds if a function's risk exceeds established limits.
Context
The core problem is that modern AI agents and assistants (such as Cursor) are highly effective at writing working code, but often do so at the expense of accumulating "dirty" code, excessive complexity, and codebase bloat. This leads to gradual project degradation (code rot), where a project becomes difficult to maintain due to uncontrolled technical debt growth.
Why It Matters for the Industry
For the industry, this represents a shift from simple code writing to automated quality management (governance-driven development). Tools of this type create a necessary layer of infrastructural guardrails, allowing companies to scale the use of AI agents in development without turning their codebase into unmanageable "spaghetti code."
Why It Matters for Users
Developers and teams can integrate AI assistants into their workflows with greater confidence, having an automatic "fuse" against quality reduction. This allows for maintaining architectural control and preventing technical debt accumulation through mechanical metric control directly within CI/CD pipelines.
Sources
Author
Look at AI, Editorial Team