We built a product around context cost and never measured our own tools

Описание и применение

В Meronq большие ответы инструментов увеличивали стоимость контекста. Внедрив подсчёт сущностей, ограничение в 32 КБ и удаление мёртвых строк, размер ответов сократился более чем на 90 %.

Задача

Tool responses were returning massive payloads (up to 84 KB) for simple sync and list operations, inflating context cost and degrading performance.

Результат

Payload sizes reduced by >90 %, eliminating unnecessary context cost and improving AI assistant performance.

Рабочий процесс

• 1. Modify shared response helper to return counts, changes, warnings, and a hint by default.
• 2. Apply a 32 000‑byte ceiling; oversized objects replaced by a notice with actual size.
• 3. Add escape hatch for explicit full‑detail requests.
• 4. Clean project_list to list only existing disk entries, removing dead rows from test suite registrations.
• 5. Update instrumentation to log response sizes for all dispatcher calls.

Доказательства

memory_sync (Unity project) before 85,061 B after 3,528 B

Источник

Return counts by default. memory_sync, project_scan and github_sync share one code path.

Источник

Put a ceiling on the shared response helper. Every result now passes a 32,000-byte limit.

Источник