@adrigm06/android-engineering-skill

Android Engineering Skill Package

A modular, production-oriented Android skill package for agentic engineering workflows.

Purpose

This repository is an open-source decision engine for high-impact Android engineering work:

• Architecture and modularization
• Compose UI state and design systems
• Gradle/build and CI strategy
• Testing strategy and confidence planning
• Performance, security, debugging, and release readiness

The package is built for multi-skill composition and deterministic conflict resolution.

Install with npx

Install a local copy with:

npx @adrigm06/android-engineering-skill

By default this creates ./android-engineering-skill in the current directory.

Optional target directory:

npx @adrigm06/android-engineering-skill ./path/to/install

Mental Model

The system uses four layers to keep policy stable and domain logic focused:

• SKILL.md: master entrypoint and routing
• AGENTS.md: runtime constraints and conflict arbitration
• skills/*/SKILL.md: domain logic and decision branching
• assets/, references/, templates/, examples/: reusable support artifacts

Cross-Skill Authority

Domain conflicts are resolved through an explicit authority model defined in AGENTS.md.

Domain	Authority	Primary responsibility
android-security	Critical / global override	Exploitability and sensitive data risk
android-architecture	Structural	Module boundaries and dependency direction
android-release-engineering	Production gate	Go/no-go, rollout, and rollback policy
android-performance	Runtime	Startup, jank, memory, ANR, battery
android-gradle-build	Build/CI	Build graph and CI constraints

Included Skills

• android-architecture
• android-compose
• android-gradle-build
• android-testing
• android-performance
• android-security
• android-code-review
• android-debugging
• android-release-engineering

Repository Structure

.
|- SKILL.md                   # Master entrypoint and routing
|- AGENTS.md                  # Runtime rules and conflict protocol
|- CONTEXT.md                 # Product philosophy and intent
|- assets/                    # Routing and evaluation assets
`- skills/
   `- <skill-name>/
      |- SKILL.md             # Domain decision logic and output contract usage
      |- references/          # Deep guidance
      |- templates/           # Reusable response artifacts
      `- examples/            # Good vs bad pattern contrasts

Output Quality Baseline

All non-trivial outputs follow the same section order:

• Context and constraints
• Decision and rationale
• Alternatives considered
• Tradeoffs
• Risks and mitigations
• Confidence and unknowns
• Cross-skill impacts
• Next implementation steps

Contributing and Evolution

• Improve decision quality and conflict determinism
• Keep skill names and trigger intent stable unless strongly justified
• Document material behavior changes in CHANGELOG.md

See CONTRIBUTING.md for contribution standards and compatibility expectations.

License

MIT. See LICENSE.