
Mobile Memory
- 23 installs
- 60 repo stars
- Updated June 14, 2026
- ahmed3elshaer/everything-claude-code-mobile
mobile-memory is a Claude Code skill that maintains a persistent memory of mobile project state (modules, dependencies, architecture, test coverage) across sessions.
About
mobile-memory is a Claude Code skill that maintains a persistent memory system for mobile development context across sessions. It records factual project state such as module layout, Gradle dependencies, architecture pattern, and test coverage so the agent can query it instead of re-asking. Developers use it at session start to reload project layout and after major changes to save updated state. It survives session breaks and compaction.
- Persists factual mobile project state (modules, deps, architecture, test coverage) across sessions
- Provides /memory-load, /memory-save, /memory-query, /memory-forget commands
- Distinct from instincts: retains binary factual state, not confidence-scored patterns
Mobile Memory by the numbers
- 23 all-time installs (skills.sh)
- Ranked #10,032 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 4, 2026 (Skillselion catalog sync)
mobile-memory capabilities & compatibility
- Capabilities
- context memory · project state · session persistence
- Use cases
- memory · documentation
What mobile-memory says it does
Persistent memory system that maintains mobile development context across sessions.
npx skills add https://github.com/ahmed3elshaer/everything-claude-code-mobile --skill mobile-memoryAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 23 |
|---|---|
| repo stars | ★ 60 |
| Last updated | June 14, 2026 |
| Repository | ahmed3elshaer/everything-claude-code-mobile ↗ |
What it does
Reload and query saved mobile project state (modules, dependencies, architecture, test coverage) so the agent keeps context across sessions.
Who is it for?
Keeping mobile project facts available to the coding agent across sessions and after context compaction
Skip if: Capturing coding patterns or heuristics (that is the instincts system, not this skill)
When should I use this skill?
Starting work on a mobile feature or after major project changes, to load or save project state
What you get
The agent can query saved project structure, dependencies, architecture, and test coverage instead of re-asking the user
By the numbers
- Tracks 5 memory types (project structure, dependencies, architecture, test coverage, compose screens)
- Retention window 30-90 days
Files
Mobile Memory Skill
Persistent memory system that maintains mobile development context across sessions.
Purpose
Unlike instincts (which capture patterns), memory retains factual project state:
- What modules exist
- What dependencies are installed
- What the architecture looks like
- What tests cover what code
This survives session breaks and compaction.
Memory Types
Project Structure Memory
Remembers your Android project layout:
{
"modules": ["app", "core:network", "feature:auth"],
"buildVariants": ["debug", "release", "staging"],
"featureModules": ["auth", "home", "profile"]
}Use when: Starting work on a new feature, need to know project layout
Dependencies Memory
Tracks all Gradle dependencies:
{
"libraries": [
{"name": "compose-runtime", "group": "androidx.compose", "version": "1.5.0"}
],
"kgpVersion": "1.9.20",
"gradleVersion": "8.2"
}Use when: Adding new dependencies, checking compatibility
Architecture Memory
Documents your architecture patterns:
{
"pattern": "mvi",
"uiLayer": {"screens": ["Home", "Profile"]},
"dataLayer": {"repositories": ["UserRepository"]},
"di": {"framework": "koin", "modules": ["appModule"]}
}Use when: Onboarding new developers, explaining codebase
Test Coverage Memory
Tracks test metrics:
{
"totalCoverage": 78,
"trend": "improving",
"failingTests": [
{"class": "AuthViewModelTest", "method": "testLogin"}
]
}Use when: Planning testing work, tracking quality goals
Compose Screens Memory
Indexes all Composable screens:
{
"screens": [
{"name": "HomeScreen", "route": "home", "file": "HomeScreen.kt"}
]
}Use when: Finding screens, understanding navigation
Usage
Load Memory
# At session start - load all memory
/memory-load all
# Load specific type
/memory-load project-structure
/memory-load dependenciesSave Memory
# Save current state
/memory-save project-structure
/memory-save test-coverage
# Save all (usually automatic)
/memory-save allQuery Memory
# Ask questions about project
/memory-query "What modules use Ktor?"
/memory-query "Which screens are not tested?"
/memory-query "What's the test coverage for auth module?"Forget Memory
# Remove stale memory
/memory-forget recent-changes
/memory-forget --older-than 90daysSummary
# Get overview of all memory
/memory-summaryMemory Refresh Triggers
Memory auto-refreshes on:
- Gradle sync: Dependencies, build variants
- File changes: Recent changes, architecture
- Test runs: Test coverage, failing tests
- Session start: Load all memory
- Session end: Save all memory
Memory vs Instincts
| Aspect | Memory | Instincts |
|---|---|---|
| Content | Factual state | Patterns |
| Examples | Module list, deps | "Use collectAsStateWithLifecycle" |
| Updates | On changes | On observations |
| Confidence | Binary (exists/doesn't) | 0.0-1.0 score |
| Retention | 30-90 days | Persistent |
Integration
With Checkpoints
Checkpoints include memory state:
{
"checkpoint": {
"memory": {
"project-structure": {...},
"dependencies": {...}
}
}
}Restoring a checkpoint restores memory too.
With Compaction
Memory survives compaction:
- Recent memory: Kept as-is
- Old memory: Summarized
- Always available via
/memory-query
With Instincts
Memory informs instinct extraction:
- Project structure → Where to look for patterns
- Dependencies → What frameworks are used
- Architecture → What patterns to expect
Best Practices
1. Let it auto-refresh: Memory updates automatically on hooks 2. Query, don't remember: Use /memory-query instead of asking user 3. Validate on load: Check memory matches actual project 4. Update after changes: Run /memory-save after major changes 5. Clean up: Use /memory-forget to remove stale data
Example Session
User: I need to add a new feature for user profiles
Agent: /memory-query project-structure
Response: Found modules: app, core:network, feature:auth
No profile module exists.
Agent: /memory-query dependencies
Response: Using Compose 1.5.0, Ktor 2.3.0, Koin 3.4.0
Agent: Based on memory, I'll create feature:profile module
following your existing architecture pattern.---
Remember: Memory is about what exists, not what should be. That's what instincts are for.
Related skills
FAQ
How is memory different from instincts?
Memory retains factual project state (module list, dependencies) with binary existence, while instincts capture confidence-scored coding patterns.
Does memory survive compaction?
Yes. Recent memory is kept as-is, old memory is summarized, and everything stays available via /memory-query.