fix_code
Evaluate code and automatically apply fixes to resolve security, performance, and quality issues.
Instructions
Evaluate code with the Judges Panel and automatically apply all available auto-fix patches. Returns the fixed code along with a summary of applied and remaining findings. Use this to fix security, performance, and quality issues in a single step.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| code | Yes | The source code to evaluate and fix. Include the full file content. | |
| config | No | Optional inline configuration (same format as .judgesrc) | |
| context | No | Optional additional context about the code — e.g., what the code does, which framework it uses. | |
| judgeId | No | Optional: restrict fixes to a single judge. One of: accessibility, agent-instructions, ai-code-safety, api-contract, api-design, authentication, backwards-compatibility, caching, ci-cd, cloud-readiness, code-structure, compliance, concurrency, configuration-management, cost-effectiveness, cybersecurity, data-security, data-sovereignty, database, dependency-health, documentation, error-handling, ethics-bias, framework-safety, hallucination-detection, iac-security, intent-alignment, internationalization, logging-privacy, logic-review, maintainability, model-fingerprint, multi-turn-coherence, observability, over-engineering, performance, portability, rate-limiting, reliability, scalability, security, software-practices, testing, ux, false-positive-review | |
| language | Yes | The programming language of the code (e.g., 'typescript', 'python', 'javascript', 'csharp', 'java'). | |
| minConfidence | No | Minimum finding confidence to include fixes for (0-1, default: 0.5) |