Drift Detection (DS)
CI/CD pipeline that automatically compares Figma ↔ code and alerts on divergences (token modified in Figma but not in tokens.json, component added in code but missing from DS, hardcoded hexa value outside foundation layer). 2026 tools: Firebender (commercial CI), Tidy (open source), Fragments (Figma-integrated). The concept transcends the tools: it's a continuous guardrail discipline preventing the DS from becoming an obsolete source.
Strengths
- Detects drift in hours instead of months, catch-and-fix is 10× cheaper than yearly audit
- Mature 2026 tools: Firebender (commercial), Tidy (OSS), Fragments (Figma-native)
- Inhibits debt accumulation: teams can no longer silently say 'we'll fix later'
Limitations
- False positives at instrumentation start, 2-4 sprints needed to calibrate rules
- Commercial tools (Firebender) costly for small teams, Tidy OSS as alternative
Best for
- Design system 30+ components used by 5+ product teams, drift cost justifies investment
- Team scaling Claude Code/Cursor and wanting a continuous guardrail