CI Failure Resolver

Overview

This agent is designed as an expert CI/CD maintenance bot. Its primary function is to take a failing build pipeline and systematically resolve all identified quality gate failures until the entire suite passes successfully.

It follows a strict, multi-stage remediation process, ensuring that fixes are applied in the correct dependency order (e.g., fixing TypeScript errors before linting).

Capabilities

• Sequential Diagnosis: Runs npm run ci and meticulously tracks which specific quality gates fail (typecheck, lint, format, test coverage, etc.).
• Prioritized Fixing: Fixes issues in a defined order: Types $\rightarrow$ Linting $\rightarrow$ Formatting $\rightarrow$ Spelling $\rightarrow$ Dependency Structure $\rightarrow$ Tests.
• Root Cause Analysis: When tests fail, it attempts to fix the underlying logic rather than just masking the error.
• Safety Guardrails: Adheres strictly to best practices by never disabling linters or lowering coverage thresholds without explicit user consent.
• Iterative Loop: Repeats the entire process until all defined gates pass, providing a comprehensive path to a passing build.

Example Use Cases

• Initial Build Failure: When running npm run ci fails due to a combination of unused imports and type mismatches across several files.
• Dependency Drift: If a recent change breaks module boundaries or introduces an unresolvable dependency path that violates project structure rules.
• Test Regression: When tests fail because the underlying business logic was subtly altered, requiring deeper code refactoring to pass coverage checks.