skills.rest
IsKenKenYa
IsKenKenYaCOMMUNITY
Explore all skills by this author →
GitHub

hmos-memory-tier-optimizer

Optimizes HarmonyOS app memory on low-end devices through tiered degradation and before-after verification.

2|Updated Jun 21, 2026
Search Tags:#performance#memory optimization#hdc#harmonyos#arkts#device tiering#low-end devices
One-click install
npx skills add https://github.com/IsKenKenYa/skills --skill hmos-memory-tier-optimizer-iskenkenya
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hmos-memory-tier-optimizer
Source: https://github.com/IsKenKenYa/skills/tree/main/skills/harmonyos/solutions/quality/performance/hmos-memory-tier-optimizer
Command: npx skills add https://github.com/IsKenKenYa/skills --skill hmos-memory-tier-optimizer-iskenkenya

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? HarmonyOS apps often exceed memory budgets on low-end devices, causing OOM crashes and jank. This Skill runs a full closed-loop workflow: collect meminfo/UI tree/screenshots via hdc, analyze bottlenecks against 14 detection rules, propose tiered optimization schemes, modify and review ArkTS/C-API code, build and install the HAP, re-measure, and produce a patch plus comparison report. ## Core Features & Use Cases - End-to-end memory optimization loop: baseline collection, data-driven analysis, user-confirmed code changes, ArkTS rule-based code review, build/install, post-optimization re-measurement, and patch/report generation. - 12 built-in optimization schemes: cachedCount tiering, image memory optimization, resource leak fixes, Web preload/prerender/offline-reuse tiering, media degradation, Tabs cache, component detaching, QuickJS limits, plus Flutter VMA and RN-specific schemes, all built on a shared device tiering mechanism (low/medium/high). - Use Case: A developer reports their HarmonyOS app exceeds PSS limits on low-end phones. The Skill collects meminfo, detects a hardcoded cachedCount of 50 in a long list, applies device-tiered cachedCount values, verifies the build, and reports a before/after PSS reduction. ## Quick Start Ask the assistant to optimize this HarmonyOS app's memory for low-end devices and produce a before-and-after PSS comparison report.

Frequently Asked Questions about hmos-memory-tier-optimizer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce HarmonyOS app memory usage on low-end devices?▼

Collect meminfo with hdc, identify dominant memory types (native heap, ArkTS heap, GL), then apply device-tiered schemes such as dynamic cachedCount, image autoResize, Web preload limits, and resource leak fixes. Verify with a before/after PSS comparison.

How to collect HarmonyOS meminfo and UI tree with hdc?▼

Use hdc shell hidumper --mem <PID> for meminfo, uitest dumpLayout for the UI tree, and uitest screenCap for screenshots. The Skill's collect_memory scripts automate this per platform and save results into a TestData directory.

What is device tiering in HarmonyOS memory optimization?▼

Device tiering classifies phones as low, medium, or high using RAM size and CPU model, optionally downgraded by cold-start latency. Optimization parameters like cachedCount and Web preload counts are then set per tier instead of hardcoded.

Does this work for Flutter or React Native HarmonyOS projects?▼

Yes, it includes Flutter-specific rules for Vulkan VMA allocator memory and RN-specific rules for background GC and image resizing. Project type is detected by scanning for src/flutter/ or @rnoh/react-native-openharmony before applying those rules.

Why does ArkTS code review require compile verification?▼

Text-only review cannot detect type errors such as bigint versus number arithmetic from APIs like hidebug.getSystemMemInfo. The Skill prefers the check_ets_files MCP, falling back to hvigorw assembleHap or manual DevEco Studio compilation.

Will the optimization modify my code automatically?▼

No. The analysis phase only outputs candidate schemes. Code changes happen only after the user confirms the scheme, affected files, and verification approach, and a git diff patch is generated for review or rollback.

Related Skills

Explore More Software Engineering Skills→
openharmonyinsightopenharmonyinsight
31

ohos-app-build-debug

Build, install, launch, and debug HarmonyOS/OpenHarmony apps via hvigorw and hdc.

Official
Advanced
Ankh-StudioAnkh-Studio

regression

Run automated regression tests comparing evaluations against baseline metrics.

Official
Advanced
ollieb89ollieb89

python-performance-optimization

Profile CPU and memory usage to optimize Python code performance.

Community
Advanced
IvanMurzakIvanMurzak
4.0k

profiler-get-memory-stats

Reads Unity Profiler memory counters and returns statistics in megabytes.

Community
Basic