Bob Truth Score: Code Reality Audit

Bob Truth Score solves a problem every engineering team recognizes: Jira says the work is Done, but the code may tell a very different story. Sprint reports, velocity charts, and manager updates all depend on ticket status, yet traditional project tools cannot understand whether the implementation actually matches the intent of the ticket. This project uses Bob as a semantic audit layer between Jira and GitHub. It reads closed Jira tickets, parses their acceptance criteria, pulls linked pull requests and full diffs, then evaluates the actual code changes against what was promised. Instead of trusting status labels, Bob produces a Truth Score from 0 to 100 and classifies each ticket as Done, Partial, or Theatrical. The scoring is based on five practical dimensions: intent alignment, acceptance criteria coverage, test presence, PR-to-ticket coherence, and regression risk. The result is not just a score, but an actionable report that explains what is missing, where the evidence is weak, and what the team should fix next. The project also includes a manager-friendly dashboard that can run audits from a JQL query, show summary counts, display ticket-level verdicts, and surface recommendations. It supports mock mode for fast demos and live Jira/GitHub mode for real teams. Beyond ticket audits, it adds bus factor analysis to identify risky areas of the codebase that depend too heavily on one developer. Bob Truth Score turns project management from theater into evidence. It gives teams a practical way to verify delivery, catch incomplete work earlier, improve engineering accountability, and protect managers from making decisions based on unreliable status updates.