Skip to content

neha/rn-developer-skills

v1.2.0MIT

Vendor-neutral React Native skills for developers and AI agents. Spec/Skill-Driven Development from spec authoring through code review.

Changelog

All notable changes to this repository are documented here. This project follows semantic versioning.

Per-skill history lives in skills/CHANGELOG.md.

Unreleased

Added

  • error-handling, notifications, theming, upgrades, and release-and-updates skills.
  • native-integration skill for permissions, native modules, platform APIs, background tasks, and listener cleanup.
  • forms-and-validation skill for form state, validation timing, keyboard handling, error focus, and submit safety.
  • observability skill for crash reporting, breadcrumbs, performance traces, analytics, and PII-safe logging.
  • i18n-and-localization skill for extracted strings, plurals, locale-aware formatting, RTL layout, and fallback locales.
  • .coderabbit.yaml so pull requests get an automated review aligned with the skill index, scripts, and workflows. Reviews stay advisory until branch protection requires the check.
  • Claude Code plugin and marketplace manifests (.claude-plugin/plugin.json, .claude-plugin/marketplace.json) so the collection can be installed with claude plugin marketplace add.
  • Root plugin.json (Agent Plugins 1.0) so Kiro can import this repository as a power from GitHub.
  • Kiro submission details: author email and URL on plugin.json, plus a privacy policy and support contact in the README.
  • Cross-platform installation scripts for Codex and Claude (scripts/install.sh, scripts/install.ps1, and scripts/install.cmd).
  • docs/skill-authoring-guide.md — worked example for adding a new skill (error-handling).
  • skills/CHANGELOG.md — per-skill version history.
  • schemas/plugin.schema.json and scripts/validate-plugin.mjs — Cursor plugin manifest validation.
  • platforms and react-native-version frontmatter fields on all skills; ## Applicability section in each skill.

Changed

  • Server-data examples use the app's existing cache. The install guide states how Cursor and Claude Code treat two copies of the same skill.
  • Reviews file a finding only on a line the diff touches. critical-rules is limited to text crashes, hooks, updates after unmount, and secrets. Each skill states what is merge-blocking. Cold start order and safe area live in architecture. Validation checks that code-review routes every skill.
  • Reviews apply critical-rules on every diff and open other skills only when the change touches them. critical-rules keeps crash and data-loss checks and points security detail at security. Layout rules follow the app's existing structure. Performance advice virtualises lists that can grow and skips blanket memoisation. Specs, the install guide, and a worked review match the 19 skills. Plugin manifests are 1.2.0.
  • Install guide counts 14 skills. code-review routes critical-rules and conventions on every change. Spec, critical-rules, conventions, and code-review include a correct/incorrect example. Plugin manifests are 1.1.0. Kiro section records the registry submission.
  • Plugin logo in .cursor-plugin/plugin.json now points to assets/logo.svg.
  • scripts/validate-skills.mjs validates skill metadata, README index, and plugin manifest.

1.0.0

Initial public release.

Added

  • spec-authoring skill teaching Spec/Skill-Driven Development (requirements → design → tasks).
  • Focused skills: architecture, performance, accessibility, testing, state-and-data, security.
  • Baseline skills: critical-rules (crash/data-loss/security/compliance safety) and conventions (structure and style).
  • code-review skill reworked as an audit entry point that routes each concern to its focused skill.
  • Standardised SKILL.md frontmatter (name, description, version, optional tags).
  • README.md with a skill index and installation instructions.
  • templates/SKILL.md and templates/feature-spec.md.
  • scripts/validate-skills.mjs to validate repository structure and skill metadata.
  • CONTRIBUTING.md and LICENSE (MIT).

Changed

  • Skills consolidated under a single canonical skills/ location.