Android Compose

Purpose

Design and review Compose UI systems with predictable state, bounded recomposition, and architecture-aligned ownership under real product constraints.

Scope and authority

This skill is the lead authority for:

• Compose UI state/event modeling
• recomposition and stability decisions
• navigation/event flow patterns inside UI layer boundaries
• composable decomposition and screen architecture

This skill must defer when applicable:

• structural boundaries to android-architecture
• runtime SLA hard constraints to android-performance
• security controls to android-security

When to use

• building new Compose screens/components
• refactoring unstable state and recomposition hotspots
• defining event/navigation flow models
• auditing Compose architecture quality
• balancing UI quality with performance and delivery constraints

Required inputs

Gather minimum decision-shaping inputs:

• screen complexity and user journey criticality
• current state holders and ownership boundaries
• jank/recomposition evidence (if performance concern exists)
• navigation complexity and cross-screen coupling
• team familiarity with Compose patterns and testing depth

If missing data blocks precision, continue with assumptions and confidence downgrade.

Decision engine workflow

1. Classify screen/system complexity.
2. Select state model branch and ownership strategy.
3. Validate composable boundaries against architecture constraints.
4. Evaluate recomposition/stability risk using available evidence.
5. Resolve runtime tradeoffs with performance constraints.
6. Define testing and preview strategy by risk.
7. Return implementation plan with fallback options.

Branching decision tree

Branch A: UI complexity

• Simple (single-flow, low concurrency):
• Prefer straightforward UiState sealed model + single state holder.
• Moderate (multiple async sources, conditional rendering):
• Use richer UiState + explicit event reducer style.
• Segment composables by responsibility and state domain.
• High (high interaction density, optimistic updates, cross-screen state):
• Use explicit intent/event model, deterministic reducers, and effect channeling.
• Tighten ownership boundaries and instrumentation.

Branch B: state ownership

• Hoist to screen-level state holder when data drives multiple child composables.
• Keep local state only for ephemeral UI concerns (focus, transient toggles, animation state).
• If local state grows domain semantics, escalate ownership upward.

Branch C: recomposition strategy

• If no jank evidence and moderate complexity:
• prefer clarity-first decomposition; avoid premature micro-optimizations.
• If measured hotspots exist:
• reduce unstable parameter propagation
• isolate expensive subtrees
• apply targeted memoization and state derivation controls

Branch D: navigation/event modeling

• keep navigation events explicit and one-directional from state holder boundary
• avoid distributed navigation logic inside unrelated leaf composables

Conflict handling and composition

When used with other skills:

• With android-performance:
• measured runtime bottlenecks may require short-term UI compromises
• document expected UX impact and reversion criteria
• With android-architecture:
• Compose convenience cannot violate layer boundaries
• reject data access shortcuts in composables
• With android-testing:
• high interaction complexity requires stronger UI+integration coverage
• With android-release-engineering:
• if release risk is high, prefer low-regression changes and stage risky refactors

Real-world tradeoff guidance

Permit context-justified compromises when explicit:

• temporary mixed View+Compose interop is acceptable during migration
• partial state normalization is acceptable under deadlines with follow-up plan
• readability-first implementation is acceptable when performance evidence is weak

Avoid presenting temporary compromises as final best practice.

Anti-pattern detection

• giant composables with mixed rendering, state mutation, and business logic
• unstable parameter graphs triggering broad recompositions
• overuse/misuse of remember or derivedStateOf without measurable benefit
• side effects hidden in render path
• navigation scattered across unrelated UI nodes

Uncertainty protocol

Always report confidence:

• High (>= 0.80)
• Medium (0.60-0.79)
• Low (< 0.60)

For medium/low confidence:

• list assumptions and what evidence is missing
• provide at least one alternative state/structure strategy
• suggest the minimum instrumentation required to choose decisively
• escalate to android-performance or android-architecture when conflict is structural/runtime

Cross-skill handoff payload

Use the standard payload defined in ../../AGENTS.md (section: Cross-skill handoff contract). Set requesting_skill to android-compose.

Output contract

Follow global order from ../../AGENTS.md:

1. Context and constraints
2. Decision and rationale
3. Alternatives considered
4. Tradeoffs
5. Risks and mitigations
6. Confidence and unknowns
7. Cross-skill impacts
8. Next implementation steps

Include Compose-specific artifacts:

• UI state model
• Composable structure
• State hoisting strategy
• Navigation/events model
• Recomposition and stability risks
• Preview/testing plan

Related resources

• references/compose-state-patterns.md
• examples/common-pitfalls.md

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: adrigm06
• Source: adrigm06/Android-Engineering-Skill
• License: MIT

Install and usage instructions live in the source repository linked above.