Why I Built Inkspect: How Do I Know the AI Is Really Done?

Описание и применение

Разработчик создал Inkspect, чтобы убедиться, что AI‑агент действительно завершил работу, проверяя наличие ошибок, доказательств и исправлений перед переходом к следующей задаче.

Задача

When a coding agent claims completion, the output may still contain hidden issues that surface later, causing regressions and wasted effort.

Результат

After integrating Inkspect, the team could detect that interrupted runs lost model IDs, add them to checkpoints, and confirm the fix with passing tests for failure, cancellation, and recovery scenarios.

Рабочий процесс

• 1. Assign a task to the coding agent.
• 2. Agent completes the task and generates a report.
• 3. Inkspect reviews the report, identifies findings, and records evidence.
• 4. If needed, Inkspect applies fixes and re‑checks the code.
• 5. Once all findings are resolved, the agent receives a handoff to start the next task.

Доказательства

An agent would finish a task, but more problems would appear when I checked the result.

Источник

The engine recorded model IDs returned by the AI service. These labels help explain which models were involved in a review.

Источник

The fix was to include the session’s recorded model IDs in each stage checkpoint.

Источник

The test checked three cases. The saved test results show the difference: ... After the fix, all cases passed.

Источник