Tightrope
Keeping a backend and frontend in sync
Project brief
PrototypeProblem
Backend and frontend contracts drift, especially when they live in different repositories.
Outcome
When a backend and frontend drift out of contract, this CLI catches it in CI (split repos or monorepo) and returns a machine-readable verdict.
Evidence
- Static, schema-sync, and runtime checks change nothing, safe to run in CI.
- Any failing check flips the verdict and can skip expensive runtime stages.
Key decisions
- Judge zone: Static, schema-sync, and runtime checks change nothing, safe to run in CI.
- Gate zone: Any failing check flips the verdict and can skip expensive runtime stages.
- Fix zone: Agents draft remediation; humans apply. Scripts judge, agents fix.
- Output: Machine-readable report with exit codes for pass, fail, and error states.
Where the drift happens
Backend and frontend contracts drift, especially when they live in different repositories. I built a CLI that catches that drift in CI and returns a machine-readable verdict.
The verification stack
Deterministic checks do the judging: schema sync, version alignment, runtime shape. They’re safe to run in CI because they change nothing.
The remediate path
When something’s off, a Cursor-SDK agent proposes the fix, but a human still applies it. Scripts judge, agents fix. It’s the same trust boundary Agent CI/CD draws at the merge level, applied here to contract checks instead.
Edit vs merge is the authority version of the same boundary.
Reading the verdict
Machine-readable report with exit codes for pass, fail, and error states.
Tooling & discipline
Portfolio graph
Maps how the work here connects. Open the Portfolio graph →
- Professional
- Personal
- Skill
- Writing