ln-622-build-auditor_skill

Overview

This skill audits build health and code-quality tooling for a codebase and produces a structured report for the coordinator. It checks compiler and linter errors, deprecation warnings, type errors, failed tests, and build configuration issues, then returns findings with severity, location, effort, and recommendations. The skill also computes a compliance score for the Build Health category and writes a single atomic report file to the specified output directory.

How this skill works

The worker parses the incoming contextStore to identify tech stack, build tool, test framework, and output_dir. It runs build checks (compile, lint, type check, tests) using the command registry for the detected tools, captures exit codes and JSON outputs where supported, and parses stderr/stdout for warnings and errors. Findings are collected with severity levels (Critical/High/Medium/Low), location, estimated effort, and remediation advice; a penalty-based scoring algorithm produces a X/10 compliance score and a markdown report is written to {output_dir}/622-build.md in a single write operation.

When to use it

• After integrating new dependencies, language upgrades, or CI pipeline changes
• Before a release or merge to main to catch build-breaking issues early
• As a scheduled quality gate in CI to enforce build health standards
• When onboarding a repository to the delivery workflow to baseline risk
• After major refactors that could introduce type or compile regressions

Best practices

• Run the worker in CI mode with non-interactive flags and a 5-minute timeout
• Provide a complete contextStore: tech_stack, build_tool, test_framework, output_dir
• Use linter/type-checker JSON output flags to improve parsing accuracy
• Do not auto-fix; use findings to create mitigation tasks and prioritise by severity
• Treat Critical compilation failures as immediate stop conditions for delivery

Example use cases

• Detecting a TypeScript tsc failure introduced by a dependency update
• Finding linter rule violations and formatting issues that block CI quality gates
• Identifying deprecated API usage flagged by compiler warnings requiring migration
• Reporting skipped or failing tests in a payments or auth module with severity
• Spotting misconfigured build scripts or incorrect asset paths that break releases