Composing Semantics Matchers logo

Composing Semantics Matchers

Community
skydoves
composing-semantics-matchers

Use this skill to build precise Compose UI test queries by composing `SemanticsMatcher` predicates with `infix and`, `infix or`, and `operator not`, plus the prebuilt filter library (`hasText`, `hasClickAction`, `isEnabled`, `isOn`, `hasTestTag`, `hasContentDescription`, `hasParent`, `hasAnyAncestor`, `hasAnyChild`, `hasAnySibling`, `hasAnyDescendant`, `hasImeAction`, `hasScrollToKeyAction`, `hasScrollToNodeAction`, `isDialog`, `isPopup`, `isRoot`, `isFocused`, `isEditable`, `isHeading`). Covers `SemanticsMatcher.expectValue` / `keyIsDefined` / `keyNotDefined` for custom semantics keys. Use when the developer asks how to find "an enabled button with text Submit", how to combine matchers, how to filter by a custom `SemanticsPropertyKey`, how to write a hierarchical predicate, or mentions `hasParent`, `hasAnyAncestor`, `expectValue`. If the developer wants one matcher instead of three chained assertions, use this skill.

Overview

Publisherskydoves
Repositoryandroid-testing-skills
Skill namecomposing-semantics-matchers
Stars
338
Forks
14
Bundled files
Instructions only
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by skydoves on GitHub. Read the source before you install it.

Installation

Install the Composing Semantics Matchers AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/skydoves/android-testing-skills/tree/main/compose/finders/composing-semantics-matchers
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/skydoves/android-testing-skills.git /tmp/android-testing-skills
mkdir -p .claude/skills
cp -r /tmp/android-testing-skills/compose/finders/composing-semantics-matchers .claude/skills/composing-semantics-matchers
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Composing Semantics Matchers in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Composing Semantics Matchers on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Composing Semantics Matchers is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Composing Semantics Matchers — One Matcher, One Error

A SemanticsMatcher is a description string plus a (SemanticsNode) -> Boolean predicate (SemanticsMatcher.kt:26). Compose ships with a rich filter library and three combinators (and, or, not) that produce a single matcher whose error message lists every clause. This skill picks the right factory, composes precise predicates, and replaces brittle assertion chains.

When to use this skill

  • The developer wants to find "an enabled Button with text Submit" or any node matching multiple constraints.
  • The developer asks how to combine hasText and hasClickAction, or how to negate a matcher with not.
  • The developer needs to filter by a custom SemanticsPropertyKey (e.g. a domain-specific role) using expectValue/keyIsDefined.
  • The developer asks about hierarchical matchers like hasParent, hasAnyAncestor, hasAnyChild, hasAnySibling, hasAnyDescendant.
  • A test chains multiple assertIs* calls and the developer wants a single, declarative predicate.

When NOT to use this skill

  • The query is "find by tag/text/content description only" — use ../finding-nodes-by-tag-text-content/SKILL.md.
  • The right node exists but only as a relative — use ../traversing-the-semantics-tree/SKILL.md.
  • The matcher is fine; only the assertions need rework — see ../../assertions/asserting-node-state-and-text/SKILL.md.

Prerequisites

  • androidx.compose.ui:ui-test on the test classpath. See ../../setup/configuring-test-dependencies/SKILL.md if unresolved.
  • Comfort with Modifier.semantics { … } in production code. Custom keys must be set via Modifier.semantics { customKey = value } to be matchable from tests.

Workflow

  • 1. Build matchers from the prebuilt filter library. Every prebuilt matcher lives in commonMain/.../Filters.kt. Group by signal:

    SignalMatchersFile:line
    Enabled / disabledisEnabled(), isNotEnabled()Filters.kt:38-47
    Toggle (Switch, Checkbox)isToggleable(), isOn(), isOff()Filters.kt:54-70
    Selection (Tab, RadioButton)isSelectable(), isSelected(), isNotSelected()Filters.kt:77-93
    FocusisFocusable(), isNotFocusable(), isFocused(), isNotFocused()Filters.kt:100-122
    Click actionhasClickAction(), hasNoClickAction()Filters.kt:129-136
    Scroll actionhasScrollAction(), hasNoScrollAction(), hasScrollToIndexAction(), hasScrollToKeyAction(), hasScrollToNodeAction()Filters.kt:143-425
    Text actions (TextField)hasSetTextAction(), hasInsertTextAtCursorAction(), hasPerformImeAction(), hasImeAction(ImeAction)Filters.kt:365-392
    Focus actionhasRequestFocusAction()Filters.kt:399
    Content descriptionhasContentDescription(value, substring, ignoreCase), hasContentDescriptionExactly(vararg)Filters.kt:165-212
    TexthasText(text, substring, ignoreCase), hasTextExactly(vararg, includeEditableText=true)Filters.kt:229-293
    A11y / state metadatahasStateDescription, hasProgressBarRangeInfo, isHeading, isDialog, isPopup, isHiddenFromAccessibility, isEditable, isRootFilters.kt:301-439
    Test taghasTestTag(testTag)Filters.kt:326-327
    HierarchicalhasParent(matcher), hasAnyChild(matcher), hasAnySibling(matcher), hasAnyAncestor(matcher), hasAnyDescendant(matcher)Filters.kt:446-530

    hasText searches SemanticsProperties.Text, SemanticsProperties.EditableText, AND SemanticsProperties.InputText simultaneously (Filters.kt:229-258). hasContentDescription matches against the merged list of descriptions (Filters.kt:165-188).

  • 2. Compose with and / or / not. Combinators come from SemanticsMatcher.kt:60-74. Each produces a new matcher whose description reads (left) && (right) / (left) || (right) / NOT (inner). The framework prints this description verbatim on assertion failure, so the failure message shows exactly which clause(s) failed.

  • 3. Use SemanticsMatcher.expectValue for custom keys. When the production code declares a custom SemanticsPropertyKey<T> and writes it via Modifier.semantics { customKey = value }, match it with SemanticsMatcher.expectValue(customKey, expectedValue) (SemanticsMatcher.kt:33-37). For "this key is set at all", use SemanticsMatcher.keyIsDefined(customKey) (SemanticsMatcher.kt:40-42); for "this key is NOT set", use keyNotDefined (SemanticsMatcher.kt:45-47). Built-in matchers like isOn() are themselves thin wrappers over expectValue(SemanticsProperties.ToggleableState, On) (Filters.kt:61-62).

  • 4. Pass the composed matcher to onNode/onAllNodes. onNode(matcher, useUnmergedTree=false) and onAllNodes(matcher, useUnmergedTree=false) are the canonical entry points (SemanticsNodeInteractionsProvider.kt). Both onNodeWithTag and friends are convenience wrappers — drop them when the predicate gets richer.

  • 5. For hierarchical assertions, prefer hasAnyAncestor/hasAnyDescendant over multiple traversal steps. Example: "find the Text inside the dialog" → onNode(hasText("Discard?") and hasAnyAncestor(isDialog())). Cleaner than onAllNodes(isDialog()).onFirst().onChildren().filter(hasText(...)).onFirst().

Patterns

Pattern: replace assertion chains with a single matcher

kotlin
// WRONG
@Test
fun submit_isEnabledClickable() {
    rule.setContent { CheckoutScreen() }
    rule.onNodeWithText("Submit")
        .assertHasClickAction()
        .assertIsEnabled()
}
// WRONG because: each assertion is a separate fetch + error path. Failure prints "no click
// action" without telling the developer the matched node is the *enabled-but-wrong* button on
// a different screen.
kotlin
// RIGHT
@Test
fun submit_isEnabledClickable() {
    rule.setContent { CheckoutScreen() }
    rule.onNode(hasText("Submit") and hasClickAction() and isEnabled())
        .assertExists()
}

The composed matcher's description prints (Text + InputText + EditableText contains 'Submit' (ignoreCase: false)) && (OnClick is defined) && (is enabled) on failure, naming every unsatisfied clause.

Pattern: hierarchical match — "the Text inside the dialog"

kotlin
// WRONG
rule.onAllNodes(isDialog())
    .onFirst()
    .onChildren()
    .filterToOne(hasText("Discard?"))
    .assertIsDisplayed()
// WRONG because: it implicitly assumes the Text is a direct child of the dialog root.
// AlertDialog wraps its body in Column/Surface, so the Text is a descendant, not a child.
kotlin
// RIGHT
rule.onNode(hasText("Discard?") and hasAnyAncestor(isDialog()))
    .assertIsDisplayed()

Pattern: negation and "exactly one"

kotlin
// "An enabled, non-toggleable Button (so: not a Switch or Checkbox) with text Submit"
rule.onNode(hasText("Submit") and isEnabled() and !isToggleable())
    .assertHasClickAction()

!matcher invokes operator fun not() (SemanticsMatcher.kt:72-74) — works on every matcher in the library.

Pattern: custom semantics key

kotlin
// production:
val PriorityKey = SemanticsPropertyKey<Int>("Priority")
var SemanticsPropertyReceiver.priority by PriorityKey

@Composable
fun TaskRow(task: Task) {
    Row(modifier = Modifier
        .testTag(TaskRowTag)
        .semantics { priority = task.priority }
    ) { /* … */ }
}

// test:
rule.onNode(hasTestTag(TaskRowTag) and SemanticsMatcher.expectValue(PriorityKey, 1))
    .assertIsDisplayed()

// "this row has a priority assigned at all":
rule.onNode(hasTestTag(TaskRowTag) and SemanticsMatcher.keyIsDefined(PriorityKey))
    .assertIsDisplayed()

Pattern: dropdown / IME-action / scroll-to-key — narrow a TextField query

kotlin
// "the editable TextField that handles Done as IME action"
rule.onNode(hasSetTextAction() and hasImeAction(ImeAction.Done))
    .performTextInput("hello")

// "any LazyColumn child reachable by key — i.e. has IndexForKey on top of ScrollToIndex"
rule.onNode(hasScrollToKeyAction()).performScrollToKey(itemKey)

hasScrollToKeyAction is the conjunction hasKey(ScrollToIndex) and hasKey(IndexForKey) (Filters.kt:415-416). hasScrollToNodeAction adds an axis-range check (Filters.kt:419-425).

Pattern: matcher description appears in error messages

kotlin
val matcher = hasText("Submit") and hasClickAction() and isEnabled()
println(matcher.description)
// (Text + InputText + EditableText contains 'Submit' (ignoreCase: false)) &&
// (OnClick is defined) && (is enabled)

The description is human-readable and used by assert(matcher) (Assertions.kt:254-267) when the matcher does not hold. Clauses MUST be readable in isolation — that is why named matchers are preferred over inline lambdas.

Mandatory rules

  • MUST use infix and / infix or / operator not to compose matchers; MUST NOT chain assertIs* calls when a single composed matcher captures the same intent.
  • MUST wrap custom semantics keys with SemanticsMatcher.expectValue(...) or SemanticsMatcher.keyIsDefined(...); MUST NOT access node.config[CustomKey] from inside an ad-hoc lambda. The factory methods produce readable error descriptions.
  • MUST prefer hasAnyAncestor(isDialog()) / hasAnyDescendant(...) over multi-step traversal when the relationship is "somewhere above/below". Cross-reference: ../traversing-the-semantics-tree/SKILL.md.
  • MUST keep the useUnmergedTree flag at its default false unless the matcher targets an inner element collapsed by merge. Skydoves hot take #2.
  • MUST NOT redefine an existing prebuilt matcher (e.g. writing a custom SemanticsMatcher("is enabled") { … }) — use isEnabled(). Custom matchers are reserved for keys that don't have a prebuilt filter.
  • PREFERRED: prefer tag-anchored matchers — hasTestTag(SubmitTag) and isEnabled() — over text-anchored. Skydoves hot take #1.

Verification

  • Every multi-condition query uses one onNode(matcher) rather than chained finders + assertions.
  • Custom key checks use SemanticsMatcher.expectValue / keyIsDefined — node.config[...] does not appear in matcher predicates.
  • ./gradlew :app:connectedDebugAndroidTest or :app:testDebugUnitTest passes.
  • Failure messages from intentionally-broken matchers print the composed description so the failing clause is identifiable.
  • No matcher is silently shadowed by a re-declared local lambda with the same name.

References

Frequently asked questions

What does the Composing Semantics Matchers AI skill do?

Use this skill to build precise Compose UI test queries by composing `SemanticsMatcher` predicates with `infix and`, `infix or`, and `operator not`, plus the prebuilt filter library (`hasText`, `hasClickAction`, `isEnabled`, `isOn`, `hasTestTag`, `hasContentDescription`, `hasParent`, `hasAnyAncestor`, `hasAnyChild`, `hasAnySibling`, `hasAnyDescendant`, `hasImeAction`, `hasScrollToKeyAction`, `hasScrollToNodeAction`, `isDialog`, `isPopup`, `isRoot`, `isFocused`, `isEditable`, `isHeading`). Covers `SemanticsMatcher.expectValue` / `keyIsDefined` / `keyNotDefined` for custom semantics keys. Use...

Why use Composing Semantics Matchers on TypingMind?

Because you install it once and use it with any model. Composing Semantics Matchers is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Composing Semantics Matchers in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/skydoves/android-testing-skills/tree/main/compose/finders/composing-semantics-matchers. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Composing Semantics Matchers?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Composing Semantics Matchers?

As many as you like. As long as a model supports skills, you can use Composing Semantics Matchers with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Composing Semantics Matchers AI skill free?

Yes. It is published on GitHub by skydoves under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇