# Vitalyvorobyev Ringgrid API Shaping > --- - Skill: `tomevault-io/vitalyvorobyev-ringgrid-api-shaping` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/vitalyvorobyev-ringgrid-api-shaping` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/vitalyvorobyev-ringgrid-api-shaping/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Integrations & APIs - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/vitalyvorobyev-ringgrid-api-shaping --- --- name: api-shaping description: Use this when designing or refactoring public APIs across workspace crates. Keeps APIs small, explicit, and stable while allowing fast internals. --------------------------------------------------------------------------------------------------------------------------------------------------------- # API shaping (lightweight) ## Aim * Small public surface * Clear data ownership (views vs owned) * Fast internals without leaking complexity ## Prefer * `Detector` structs that own scratch buffers → avoid allocations per call. * `Config` structs with safe defaults, but don’t hide “magic” thresholds. * `ImageView` / `ImageViewMut` in APIs; keep crates buffer-agnostic. * Separate “core algorithm” from “pipeline convenience wrapper”. ## Avoid * Generic abstractions that obscure hot loops. * Exposing internal scratch buffers in public API. * “One mega function” that does everything. ## Patterns that work here * `detect_*(&mut self, img: &ImageView<_>, cfg: &Config) -> Output` * `Output` types that can be iterated cheaply (`Vec`, `Vec`) * Optional features for parallelism/SIMD later, not in baseline. --- > Converted and distributed by [TomeVault](https://tomevault.io/claim/vitalyvorobyev) — claim your Tome and manage your conversions.