bladeofgod

bladeofgod/flutter-ai-harness

Skill Claude CodeCodex

A read-only security review for a defined code scope, checking trust boundaries, sensitive data, outside input, dependencies, and changes to agent abilities.

bladeofgod/flutter-ai-harness

Skill Claude CodeCodex

适用：编写或审查 iOS Swift Native Module、Bridge Adapter、Host 接线、Info.plist/Entitlements、SwiftPM 和原生 UI。不适用：Kotlin/Android、Dart Client 或 Wire Contract 设计。触发词：Swift、iOS、Xcode、SwiftPM、Info.plist、Entitlements、actor、Sendable、AVFoundation。.

bladeofgod/flutter-ai-harness

Skill Claude CodeCodex

A testing guide for Dart and Flutter projects, covering unit, widget, route, integration, fake, and Mocktail-based tests.

bladeofgod/flutter-ai-harness

Skill Claude CodeCodex

A tool for writing, checking, or carrying out a Version 1 specification of how a user interface should behave, based on approved tasks, product rules, prototypes, or Figma designs.

bladeofgod/flutter-ai-harness

Agent Claude Code

An Android engineering role for native modules, bridge adapters, Android user interfaces, Gradle setup, platform tests, and builds. Android is Google's mobile operating system; a bridge adapter connects Android code with another runtime such as Flutter.

bladeofgod/flutter-ai-harness

Agent Claude Code

A runtime UI verification agent for apps using Marionette, a tool for connecting to and controlling an app during development. It follows an already approved behavior specification and writes a run report.

bladeofgod/flutter-ai-harness

Agent Claude Code

A planning agent that turns product requirements, design, protocols, and code evidence into an ordered technical plan and task cards. It stops before writing application code.

bladeofgod/flutter-ai-harness

Agent Claude Code

An audit and optimization agent for application assets such as images, fonts, icons, animations, and other packaged resources. It checks their usage, format, size, duplication, licensing, and effect on the built app.

bladeofgod/flutter-ai-harness

Agent Claude Code

An integration role for defining the contract between Flutter, Android, and iOS parts of an application. It specifies the messages, data fields, errors, versions, and lifecycle rules those parts use to communicate.

bladeofgod/flutter-ai-harness

Agent Claude Code

A read-only code review role for Flutter, Dart, native-platform, bridge, integration, and quality-gate changes. It checks correctness, architecture, resource lifetimes, and missing tests.

bladeofgod/flutter-ai-harness

Agent Claude Code

A read-only review role for shared capability and wire contracts, documentation, plans, task cards, and their verification evidence. A wire contract is the structured format used when separate parts of a system exchange data.

bladeofgod/flutter-ai-harness

Agent Claude Code

A read-only review role for the coding-agent system around commands, agents, skills, tasks, validators, test fixtures, evidence, and archived reports. A fixture is a prepared test case, while evidence is the recorded output proving a check ran.

bladeofgod/flutter-ai-harness

Agent Claude Code

An agent focused on implementing and reviewing native iOS modules, bridge adapters, native UI and Apple build connections using Swift and Xcode.

bladeofgod/flutter-ai-harness

Agent Claude Code

An independent security reviewer for specified tasks and code changes. It examines trust boundaries, sensitive data, external input, dependencies, and changes to an agent's capabilities, then reports evidence-based findings.

bladeofgod/flutter-ai-harness

Agent Claude Code

An independent auditor that compares a structured UI behavior specification with the production implementation. It marks each step, assertion, and teardown as covered, missing, or wrong using concrete file-and-line evidence.

bladeofgod/flutter-ai-harness

Agent Claude Code

An agent that turns approved tasks, product rules, or prototypes into a reviewable description of how a mobile app's interface should behave.

bladeofgod/flutter-ai-harness

Agent Claude Code

An implementation role that carries out one existing task card, including the requested code, tests, and validation. It is intended for Flutter, Dart, Android, iOS, bridge, and related project work within the task's stated scope.

bladeofgod/flutter-ai-harness

Agent Claude Code

A coding agent for writing Dart tests, Flutter interface tests, and integration tests for behavior that already exists.

bladeofgod/flutter-ai-harness

Command Claude Code

A release-readiness command that checks whether a code repository meets its rules for a new version, without publishing anything.

bladeofgod/flutter-ai-harness

Command Claude Code

A command for carrying out existing task cards through implementation, testing, review, fixes, and archiving. A task card is a structured file describing work to be done.

bladeofgod/flutter-ai-harness

Command Claude Code

A manually started command for auditing and running a ready UI behaviour specification on a specified Android or iOS app. It checks the implementation against the specification without changing production code.

bladeofgod/flutter-ai-harness

Command Claude Code

A command for fixing problems found in an existing code review report and then checking the changes with a new review.

bladeofgod/flutter-ai-harness

Command Claude Code

A planning command that turns one or more Figma designs—visual interface mockups—into structured engineering task cards.

bladeofgod/flutter-ai-harness

Command Claude Code

A command that creates a written specification for UI automation, meaning automated checks of how a user interface behaves. It uses confirmed task or design inputs and records the result as a YAML file.