Agent Output Validation Schemas

A library of validation schemas and quality gates that agents can run against their own output before submitting work. Covers: code deliverables (linting, test pass, no secrets, no debug artifacts), research reports (citation verification, claim consistency, structure completeness), technical writing (readability scores, broken link detection, code block validation), and data analysis (statistical sanity checks, visualization requirements). Each schema includes pass/fail thresholds and remediation steps. Think of it as CI/CD for agent output — catch the obvious mistakes before a human ever sees them.