task-verification-pipeline

A verification pipeline runs multiple checks—linting, type checking, testing, and building—in sequence to catch issues before CI. Run ./verify.sh --ui=false to execute the baseline pipeline, fix any failing stage, and re-run until all checks pass.

E2E modes control test scope: smoke runs minimal checks for rapid feedback, standard runs the full test suite, and full includes extended tests. Use --e2e=smoke|standard|full with ./verify.sh --ui=false to select the tier matching your debugging needs.

Yes. The pipeline relies on root scripts (lint, typecheck, test, build) for targeted iteration. Run specific stages independently to isolate failures, then use /verify-lint or /lint-and-learn slash commands to accelerate workflows and fix issues faster.

Run ./verify.sh --ui=false to identify the failing stage, fix that stage using the root scripts, then re-run the full pipeline. Use slash commands like /verify to re-trigger verification or /lint-and-learn to debug formatting issues quickly.

Check your repo's verification runner documentation for the canonical setup. The Skill works with any verification runner that supports stage selection and E2E tier toggling; follow your repo's documented entry point and command syntax.

The pipeline is designed for local debugging and iteration. Use ./verify.sh --ui=false or equivalent commands locally to fix failures before pushing; the same stages run in CI, ensuring consistency between your environment and the CI system.