Text prompts

Give a coding agent a bounded feature task with verification and a clean handoff.

Turn a failing check into a reproducible, tested repair instead of a speculative patch.

Guide an operations agent through evidence gathering, mitigation, repair, and verification.

Review a change for concrete correctness, security, and regression risks.

Diagnose an exception from evidence and propose the smallest verifiable fix.

Improve maintainability while explicitly preserving behavior and public contracts.

Pressure-test an HTTP or SDK contract before it becomes expensive to change.

Generate focused tests that cover boundaries, malformed input, and regressions.

Produce a current decision brief with primary sources and explicit uncertainty.

Translate a research paper into testable engineering work without overstating its claims.

Compare vendors on evidence, switching cost, operational fit, and total cost.

Turn a fuzzy quality goal into a repeatable evaluation with failure examples.

Clarify why a holistic agent cloud is different from a model gateway or app host.

Create a truthful comparison page that respects the specialist and explains fit.

Plan a useful search cluster with distinct intent instead of thin templated pages.

Turn messy interview notes into evidence-weighted themes and product decisions.