Audit workflow
The audit follows a repeatable workflow from document scanning through to re-audit. Each pass extracts claims, cross-checks them against canonical sources, and records findings with assigned owners and due dates.Terminology
Canonical component names and glossary terms are locked. All pages must use the same spelling and capitalization.Architecture
Verify that every component description matches the approved architecture diagrams and component pages.API
Verify that API signatures, request/response schemas, and endpoint paths documented in interface pages match the actual implementation.Database
Verify that schema references, table names, and field descriptions in docs match the actual database design.Task and document alignment
Verify that every task in the project management pages has a corresponding documentation page, and vice versa.Phase and task alignment
Verify that tasks are assigned to the correct phase and that phase descriptions match the roadmap.Dependency correctness
Verify that documented dependencies between components, tasks, and phases are accurate and acyclic.Research claim correctness
Verify that all research citations, algorithm descriptions, and performance claims are accurate and sourced.Mathematical consistency
Verify that equations, variable names, and notation are consistent across all math pages.Benchmark consistency
Verify that benchmark names, metrics, and baselines are consistent across evaluation and research pages.Security consistency
Verify that security pages use consistent terminology for authentication, authorization, signing, and audit logging.Versioning consistency
Verify that version numbers, release names, and compatibility statements are consistent.Overall status
The terminology section is fully locked and verified. All other sections require cross-checking against source code, database schemas, API specifications, and research papers. Schedule the first corrective action sprint to begin within one week of report generation.