adding-a-feature
Skill0 starsChanged 4 days agoMITView on GitHub
The checklist for adding one feature to this app. Check the scope, then cover the screen, the action and its test, the regenerated action reference and the screen-context entry. Use when adding or changing a screen, route or capability.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- BuilderIO/agent-nativetheir own version, first seen Oct 8, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- BuilderIO/agent-native★ 7.6ksame 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 repository29 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- accessibilitySkill
- actionsSkill
- agent-workflowSkill
- app-contextSkill
- approval-and-safetySkill
- backend-proxySkill
- background-workSkill
- build-an-appSkill
- capture-learningsSkill
- clean-architectureSkill
- customizing-the-assistantSkill
- delegate-to-agentSkill
- edit-boundariesSkill
- i18n-and-rtlSkill
- model-providersSkill
- performanceSkill
- permissions-and-platformSkill
- read-the-docsSkill
- reliable-mutationsSkill
- routingSkill
- secretsSkill
- state-managementSkill
- storing-dataSkill
- testingSkill
- ui-designSkill
- upgrade-the-frameworkSkill
- verifySkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 1 is distinctive, and 3 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.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Don't load this for starting a whole app from the template.
- Distinctive: line 12
Check the scope first. Then cover four areas, all of them, in one change:
- Very specific: line 29
1. Screen: Read through watch…/get…; change data only with registry.run(action, input), catching ActionFailure as …
- Very specific: line 35
4. Screen context: Add the route's entry to buildScreenContexts in lib/ai/screen_contexts.dart.
≈ “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
Adding a feature
Don't load this for starting a whole app from the template; see build-an-app.
Rule
Check the scope first. Then cover four areas, all of them, in one change:
- the screen;
- the action and its test;
- the regenerated action reference (and an
AGENTS.mdrow only for a primary action); - the screen-context entry for every new route.
Enforced by test/ai/parity_test.dart (an action without a test file, a stale reference, a route without a screen entry, a screen entry matching no route, or an assistant-callable action without an action_<name> label in every language file fails it) and by the lint rules (ui_calls_actions, presentation_layer_boundaries, domain_layer_boundaries, no_error_returns_in_actions, no_hardcoded_strings, no_hardcoded_colors, no_navigator_push), which run in dart analyze and tool/check.sh. Whether the scope check was done is not checked by anything.
Why
A feature that only has a screen cannot be used by the assistant. A feature with an action but no test, or a route with no context entry, leaves the assistant blind or unverified. The four areas keep the UI and the assistant equal.
How
Scope check. Does it change data the app keeps? It needs an action. Does it add a screen? It needs a route and a context entry. Does it add a table or a column? Only new tables or nullable columns, raise schemaVersion, add a schema dump and a to_v<N>_test.dart. Is it bigger than asked? Cut it and say so.
1. Screen. Read through watch…/get…; change data only with registry.run(action, input), catching ActionFailure as lib/presentation/home/home_screen.dart does. Strings from every file in assets/lang/, colors from the theme, navigation through GoRouter. Add the route in lib/utils/routes/app_router.dart.
2. Action and test. See actions. Put it in lib/domain/actions/, register it in AppActions, and write test/actions/<action_name>_test.dart by copying test/actions/create_item_test.dart. Add an action_<action_name> key to every file in assets/lang/ with a short label ("Add an item"); the chat shows it on tool rows and approval cards.
3. Reference. dart run tool/action.dart --write-reference. Add a row to the Primary actions table in AGENTS.md only for a primary action (one the person would ask for by name), not for every helper.
4. Screen context. Add the route's entry to buildScreenContexts in lib/ai/screen_contexts.dart; see app-context. If the assistant should know about it, mention it in assets/assistant/instructions.md.
Anti-patterns
- A screen that calls a repository's write methods directly.
- An action registered but untested, or a reference file edited by hand.
- A new route with no entry in
lib/ai/screen_contexts.dart. - A table change without a schema dump and migration test.
- Adding an
AGENTS.mdrow for every action. - Adding a feature nobody asked for.
Verification
dart run tool/action.dart --write-reference, thenflutter test test/ai/parity_test.dart.dart analyze --fatal-infosfrom the repo root (notflutter analyze).tool/check.shends withAll checks passed.
Related skills
actions— writing and registering the action.app-context— the screen-context entry.approval-and-safety— if the feature deletes or changes much at once.build-an-app— when the task is a whole app rather than one feature.routing— when the feature adds a route.testing— the tests for each area.verify— the last step: run the gate.
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.

