jev-quality-gate
An AI-based code quality review tool that uses the Jev model as its backend. It performs semantic code review on git diffs through YAML-defined policies, CI integration, and incremental codebase exploration.
Timeline (last 90 days)
jev-quality-gate is an AI-based code quality review tool that uses the Jev model as its backend and decision engine.
The jev-quality-gate project has been open-sourced on GitHub, with the repository located at krisitown/jev-quality-gate.
The team defines their own code specification policies in YAML format.
jev-quality-gate runs as part of the CI pipeline (and can also be run locally), fetching all git diff code blocks and querying Jev one by one as to whether each change complies with each policy.
The author used a locally deployed Qwen3.8 Flash Next model to generate code for testing the initially defined clean code policies for calibration purposes.
The jev-quality-gate project is still in an early exploratory stage, rather than a mature solution validated at large scale.
jev-quality-gate replaces dedicated policy languages such as Rego with natural-language policies described in YAML, lowering the barrier to writing policies but introducing uncertainty in how the LLM interprets them.
All Facts (7)
jev-quality-gate is an AI-based code quality review tool that uses the Jev model as its backend and decision engine.
50%UnverifiedThe jev-quality-gate project has been open-sourced on GitHub, with the repository located at krisitown/jev-quality-gate.
50%UnverifiedThe team defines their own code specification policies in YAML format.
50%Unverifiedjev-quality-gate runs as part of the CI pipeline (and can also be run locally), fetching all git diff code blocks and querying Jev one by one as to whether each change complies with each policy.
50%UnverifiedThe author used a locally deployed Qwen3.8 Flash Next model to generate code for testing the initially defined clean code policies for calibration purposes.
50%UnverifiedThe jev-quality-gate project is still in an early exploratory stage, rather than a mature solution validated at large scale.
50%Unverifiedjev-quality-gate replaces dedicated policy languages such as Rego with natural-language policies described in YAML, lowering the barrier to writing policies but introducing uncertainty in how the LLM interprets them.
50%