flutter-app-engineering
Skill0 starsChanged 13 days agoMITView on GitHub
Build, review, debug, or improve Flutter applications, including Dart implementation, state and widget lifecycles, API efficiency, navigation, secure sessions, smooth UX, native integration, and release reliability. Use for meaningful Flutter app engineering; skip standalone Dart backend/CLI work, unrelated frameworks, and trivial copy edits.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- anandartin/celestial-skillstheir own version, first seen Nov 1, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- anandartin/celestial-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- api-design-engineeringSkill
- deep-readonly-system-auditSkill
- distributed-systems-reliabilitySkill
- engineering-research-intelligenceSkill
- go-backend-engineeringSkill
- multi-tenant-access-controlSkill
- nextjs-fullstack-engineeringSkill
- nodejs-backend-engineeringSkill
- postgresql-data-engineeringSkill
- react-vite-dashboard-engineeringSkill
- real-world-test-engineeringSkill
- redis-production-engineeringSkill
- repository-structure-engineeringSkill
- saas-security-engineeringSkill
- task-subagent-orchestrationSkill
- universal-database-engineeringSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Read applicable instructions and trace the affected screen/action through state, data access, backend, persistence, and native integration.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Flutter App Engineering
Improve the requested user workflow and the engineering that supports it. Adapt to the product, supported platforms, existing architecture, and actual failure. Preserve sound conventions; explicit user requirements and repository instructions take precedence over generic recommendations.
Establish the app and workflow context
Read applicable instructions and trace the affected screen/action through state, data access, backend, persistence, and native integration. Inspect the relevant entrypoint, widget/controller/provider ownership, router, clients, tests, and platform configuration. Determine:
- The task: feature, focused repair, duplicate requests, loading/navigation, security, offline/native lifecycle, review, or measured optimization.
- Flutter/Dart versions and pins,
pubspec.yamlconstraints, lockfile, analyzer rules, generated code and target platforms/flavors. - Existing state management, dependency scopes, source of server data, storage, auth/session model, and API contracts.
- The affected cold/warm launch, route/tab transition, mutation, background/resume, notification, or restart sequence.
- Data sensitivity, account/tenant scope, freshness, replay/conflict requirements, supported devices, and evidence already available.
Bound investigation to the task. Read installed package APIs, matching Flutter/Dart docs, and relevant OS/plugin requirements before relying on changing behavior. Flutter's bundled Dart version and a standalone Dart installation may differ. Do not upgrade SDKs, change state libraries, add platforms, or restructure the whole app to fix a local problem.
Load only relevant detail
| Task or decision | Reference |
| Dart implementation, widget/element lifecycle, package structure, state/dependency ownership | Architecture and lifecycle |
| Duplicate requests, client orchestration, mutations, caching, offline persistence/sync | Data and offline behavior |
| Routing, auth restoration, deep links, account/tenant transitions, client security | Navigation and security |
| Loading/recovery, forms/lists, adaptive layout, accessibility and localization | UX and accessibility |
| Jank, startup, rendering, images, memory, isolates and profiling | Performance |
| Android/iOS lifecycle, permissions, push, real-time and background execution | Platforms and events |
| Diagnostics, focused tests, native/build compatibility and release operations | Verification and release |
| Versions, official upstream skills/tooling and authoritative research | Source and tooling register |
Use installed compatible Flutter/Dart/package skills for a genuinely matching specialized workflow, without duplicating their recipes or adopting their examples as universal architecture. Discover actual MCP capabilities when connected; otherwise use existing SDK, DevTools and project tools. Do not install optional plugins, modify agent configuration, or add test-driver hooks automatically.
This skill owns Flutter client behavior and native integration. API/backend/database specialist skills can help at their actual boundaries; none is required. UI permissions are usability measures: the backend must independently authorize sensitive data and operations.
Make the smallest sound change
State the root problem or required invariant and selected approach before editing. Reuse existing widgets, state owners, HTTP clients, routing, storage, validation and observability. Separate substantial data/business work from rendering without imposing a universal folder tree or layer per simple operation.
For changed async work, identify who starts it, which input/identity it belongs to, how stale results are rejected, and who cancels/disposes resources. A rebuild is not a reason to fetch; mounted is not freshness; a timeout is not cancellation; background execution is not durable delivery.
Represent loading, refresh, empty, filtered-empty, denied, failed and uncertain-mutation states accurately. Preserve useful content and recoverable input. Avoid adding global caches, keep-alives, isolates, polling, shimmer, or a library migration merely to conceal incorrect ownership or state transitions.
Verify and finish proportionately
Use existing focused checks for the changed behavior: formatting/analyzer, unit/widget tests, integration/device checks, or the relevant build. Choose realistic failure sequences from the actual flow rather than a universal suite. Documentation-only skill work needs normal validation, not an application fixture or agent evaluation campaign.
Claims about reduced requests need network/backend evidence and correct freshness/isolation. Claims about smoothness/performance need relevant interactions and representative profile/release conditions. State clearly when devices, platform toolchains or optional tools were unavailable.
Review the diff for compatibility, resource leaks, stale/private state, duplicate systems and unrelated changes. Report the resulting behavior, checks actually executed, observations, and material limitations; compilation alone does not establish a reliable mobile experience.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

