Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Using Strong Skipping Correctly · skydoves | DecimalAI

Frontendby skydoves

Using Strong Skipping Correctly

Use this skill to reason about Jetpack Compose's Strong Skipping Mode — the default since Kotlin 2.0.20 — including what it changes about skippability, when it does and does not auto-`remember` lambdas, and which escape hatches (`@DontMemoize`, `@NonSkippableComposable`, `@NonRestartableComposable`, `@ReadOnlyComposable`) apply where. Covers verifying the mode is active, auditing lambda capture sites, and the gaps where strong skipping does not memoize (`LazyListScope.items {}`, `Modifier.pointe

Never made a case worseMaintained

SkillScore

SkillScore95/100

Test results 95

Safety

3/3 safety checks passed

Code scan
Intent
Content

What it changes

45%95%of test cases passed +50 pts

Without skill10 of 22

With skill21 of 22

11 fixed10 passed either way1 still failing0 broke

Measured on gemini-3.6-flash8/19/2026v1

Track record

Agent runs: 12; All time

See one task, with and without the skill

One test taskWe are running Kotlin 2.0.20 in our Android project, but we need to temporarily turn off automatic strong skipping mode in our build configuration to isolate a compiler regression. How should we configure the modern Compose compiler Gradle extension to disable strong skipping?

Without the skillMissed

With the skillPassed

Same model, same task. Answers are capped and redacted.

Use this skill

Your agent runs on

Adds every file the skill has, pinned so you can update it. No account needed.

pip install "decimalai>=0.15"
decimalai skills add skydoves-using-strong-skipping-correctly
claude --plugin-dir "$(git rev-parse --show-toplevel 2>/dev/null || pwd)/.agents"

The same install from Python, for a notebook or a setup script.

Install

pip install "decimalai>=0.15"

Add the skill

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

print("Added. Start Claude Code with the command below.")

Start Claude Code with it

claude --plugin-dir "$(git rev-parse --show-toplevel 2>/dev/null || pwd)/.agents"

Type these in a Claude Code session. No account or API key. It writes the skill and any files it bundles, and asks first when its safety rating isn't Passed.

Add the DecimalAI plugin, once

/plugin marketplace add decimal-labs/decimalai-claude-plugins
/plugin install decimalai@decimalai

Install this skill

/decimalai:install skydoves-using-strong-skipping-correctly

The plugin needs python3 and curl on your machine.

No Python or account. Claude Code picks it up in its next session.

Save it into your project

curl -fsSL --create-dirs -o .claude/skills/skydoves-using-strong-skipping-correctly/SKILL.md https://app.decimal.ai/s/skydoves-using-strong-skipping-correctly/SKILL.md

Copy SKILL.md

Save it in your project, as

.claude/skills/skydoves-using-strong-skipping-correctly/SKILL.md

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Claude Code docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[claude-agent-sdk]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

import asyncio

from claude_agent_sdk import ClaudeAgentOptions
from decimalai.claude_agent_sdk import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

async def main():
    question = "Hello!"
    async with support.run(question) as run:  # one user message, one run
        async for message in run.query(options=ClaudeAgentOptions(max_turns=6)):
            print(message)

# In a notebook or async code: await main()
asyncio.run(main())

Prefer its own loader? ClaudeAgentOptions(plugins=[skills_plugin()], skills="all") reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

The same install from Python, for a notebook or a setup script. agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[claude-agent-sdk]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

import asyncio

from claude_agent_sdk import ClaudeAgentOptions
from decimalai.claude_agent_sdk import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

async def main():
    question = "Hello!"
    async with support.run(question) as run:  # one user message, one run
        async for message in run.query(options=ClaudeAgentOptions(max_turns=6)):
            print(message)

# In a notebook or async code: await main()
asyncio.run(main())

Prefer its own loader? ClaudeAgentOptions(plugins=[skills_plugin()], skills="all") reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from claude_agent_sdk import ClaudeAgentOptions, query

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
options = ClaudeAgentOptions(system_prompt=f"You are a helpful agent.\n\n{skill}", max_turns=6)

async def main():
    async for message in query(prompt="Hello!", options=options):
        print(message)

# In a notebook or async code: await main()
asyncio.run(main())

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Claude Agent SDK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[openai-agents]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

from agents import Agent, Runner
from decimalai.openai_agents import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    name=support.name,
    instructions=support.instructions,  # your prompt, then the skill list
    tools=[*support.tools],             # your tools, then load_skill
)
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

Using SandboxAgent? capabilities=[Shell(), decimalai.openai_agents.sandbox_skills()] points its own loader at the same folder.

Install

pip install "decimalai[openai-agents]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

from agents import Agent, Runner
from decimalai.openai_agents import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    name=support.name,
    instructions=support.instructions,  # your prompt, then the skill list
    tools=[*support.tools],             # your tools, then load_skill
)
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

Using SandboxAgent? capabilities=[Shell(), decimalai.openai_agents.sandbox_skills()] points its own loader at the same folder.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from agents import Agent, Runner

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = Agent(name="support", instructions=f"You are a helpful agent.\n\n{skill}")
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

OpenAI Agents docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[openai]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

import json

from openai import OpenAI
from decimalai.openai import agent_config

client = OpenAI()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your tools, as Chat Completions takes them, and what answers each.
TOOLS = [{"type": "function", "function": {
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters": {"type": "object", "properties": {"order_id": {"type": "string"}},
                   "required": ["order_id"]},
}}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    # run.system is your prompt, then the skill list.
    messages = [{"role": "system", "content": run.system}, run.message]
    while True:
        reply = client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=messages,
            tools=[*TOOLS, *run.tools],  # your tools, then load_skill
        ).choices[0].message
        messages.append(reply)
        if not reply.tool_calls:
            break
        for call in reply.tool_calls:
            # load_skill is answered by the run; your own tools by FUNCTIONS.
            function = FUNCTIONS.get(call.function.name)
            messages.append(run.tool_result(call) or {
                "role": "tool", "tool_call_id": call.id,
                "content": function(**json.loads(call.function.arguments)) if function
                else f"There is no tool named {call.function.name!r}.",
            })
print(reply.content)

# The Responses API takes run.response_tools instead of run.tools.

Install

pip install "decimalai[openai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

import json

from openai import OpenAI
from decimalai.openai import agent_config

client = OpenAI()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your tools, as Chat Completions takes them, and what answers each.
TOOLS = [{"type": "function", "function": {
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters": {"type": "object", "properties": {"order_id": {"type": "string"}},
                   "required": ["order_id"]},
}}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    # run.system is your prompt, then the skill list.
    messages = [{"role": "system", "content": run.system}, run.message]
    while True:
        reply = client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=messages,
            tools=[*TOOLS, *run.tools],  # your tools, then load_skill
        ).choices[0].message
        messages.append(reply)
        if not reply.tool_calls:
            break
        for call in reply.tool_calls:
            # load_skill is answered by the run; your own tools by FUNCTIONS.
            function = FUNCTIONS.get(call.function.name)
            messages.append(run.tool_result(call) or {
                "role": "tool", "tool_call_id": call.id,
                "content": function(**json.loads(call.function.arguments)) if function
                else f"There is no tool named {call.function.name!r}.",
            })
print(reply.content)

# The Responses API takes run.response_tools instead of run.tools.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from openai import OpenAI

client = OpenAI()
skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
reply = client.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[
        {"role": "system", "content": f"You are a helpful agent.\n\n{skill}"},
        {"role": "user", "content": "Hello!"},
    ],
).choices[0].message

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

OpenAI SDK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[anthropic]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

import anthropic
from decimalai.anthropic import agent_config

client = anthropic.Anthropic()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

question = "Hello!"
with support.run(question) as run:  # one user message, one run
    answer = client.beta.messages.tool_runner(
        model="claude-haiku-5-5",
        max_tokens=4096,
        system=run.system,     # your prompt, then the skill list
        tools=[*run.tools],    # your tools, then load_skill
        messages=[{"role": "user", "content": question}],
    ).until_done()

Install

pip install "decimalai[anthropic]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

import anthropic
from decimalai.anthropic import agent_config

client = anthropic.Anthropic()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

question = "Hello!"
with support.run(question) as run:  # one user message, one run
    answer = client.beta.messages.tool_runner(
        model="claude-haiku-5-5",
        max_tokens=4096,
        system=run.system,     # your prompt, then the skill list
        tools=[*run.tools],    # your tools, then load_skill
        messages=[{"role": "user", "content": question}],
    ).until_done()

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
import anthropic

client = anthropic.Anthropic()
skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
reply = client.messages.create(
    model="claude-haiku-5-5",
    max_tokens=4096,
    system=f"You are a helpful agent.\n\n{skill}",
    messages=[{"role": "user", "content": "Hello!"}],
)

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Anthropic docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[google-genai]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

from google import genai
from google.genai import types
from decimalai.google_genai import agent_config

client = genai.Client()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your functions, as declarations for the model, and what answers each.
TOOLS = [{"function_declarations": [{
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters_json_schema": {"type": "object", "properties": {"order_id": {"type": "string"}},
                               "required": ["order_id"]},
}]}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    config = types.GenerateContentConfig(
        system_instruction=run.system,     # your prompt, then the skill list
        tools=[*TOOLS, *run.tool_params],  # your functions, then load_skill
    )
    contents = [{"role": "user", "parts": [{"text": run.message}]}]
    while True:
        response = client.models.generate_content(
            model="gemini-3.5-flash", contents=contents, config=config,
        )
        if not response.function_calls:
            break
        contents.append(response.candidates[0].content)
        parts = []
        for call in response.function_calls:
            # load_skill is answered by the run; your own functions by FUNCTIONS.
            # A loop of your own rather than automatic function calling, which
            # raises KeyError when the model asks for a function you did not pass.
            function = FUNCTIONS.get(call.name)
            parts.append(run.tool_result(call) or {"function_response": {
                "name": call.name, "id": call.id,
                "response": {"result": function(**(call.args or {})) if function
                             else f"There is no function named {call.name!r}."},
            }})
        contents.append({"role": "user", "parts": parts})
print(response.text)

Install

pip install "decimalai[google-genai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

from google import genai
from google.genai import types
from decimalai.google_genai import agent_config

client = genai.Client()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your functions, as declarations for the model, and what answers each.
TOOLS = [{"function_declarations": [{
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters_json_schema": {"type": "object", "properties": {"order_id": {"type": "string"}},
                               "required": ["order_id"]},
}]}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    config = types.GenerateContentConfig(
        system_instruction=run.system,     # your prompt, then the skill list
        tools=[*TOOLS, *run.tool_params],  # your functions, then load_skill
    )
    contents = [{"role": "user", "parts": [{"text": run.message}]}]
    while True:
        response = client.models.generate_content(
            model="gemini-3.5-flash", contents=contents, config=config,
        )
        if not response.function_calls:
            break
        contents.append(response.candidates[0].content)
        parts = []
        for call in response.function_calls:
            # load_skill is answered by the run; your own functions by FUNCTIONS.
            # A loop of your own rather than automatic function calling, which
            # raises KeyError when the model asks for a function you did not pass.
            function = FUNCTIONS.get(call.name)
            parts.append(run.tool_result(call) or {"function_response": {
                "name": call.name, "id": call.id,
                "response": {"result": function(**(call.args or {})) if function
                             else f"There is no function named {call.name!r}."},
            }})
        contents.append({"role": "user", "parts": parts})
print(response.text)

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from google import genai
from google.genai import types

client = genai.Client()
skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
response = client.models.generate_content(
    model="gemini-3.5-flash",
    contents="Hello!",
    config=types.GenerateContentConfig(system_instruction=f"You are a helpful agent.\n\n{skill}"),
)

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Gemini API docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[langchain]" langchain-openai
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

from langchain.agents import create_agent
from decimalai.langchain import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = create_agent(
    "openai:gpt-5.4-mini",
    tools=[],                         # your tools; the middleware adds load_skill
    middleware=[support.middleware],  # your prompt, the skill list and load_skill
)
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

Install

pip install "decimalai[langchain]" langchain-openai

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

from langchain.agents import create_agent
from decimalai.langchain import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = create_agent(
    "openai:gpt-5.4-mini",
    tools=[],                         # your tools; the middleware adds load_skill
    middleware=[support.middleware],  # your prompt, the skill list and load_skill
)
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from langchain.agents import create_agent

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = create_agent("openai:gpt-5.4-mini", tools=[], system_prompt=f"You are a helpful agent.\n\n{skill}")
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

LangChain docs

CrewAI's own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" "crewai>=1.15.27"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

import decimalai.crewai
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                               # your own CrewAI LLM
    skills=decimalai.crewai.skills(),      # the installed folder
    tools=decimalai.crewai.skill_tools(),  # reads a skill's files, runs its scripts
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

The same install from Python, for a notebook or a setup script. CrewAI's own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" "crewai>=1.15.27"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

import decimalai.crewai
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                               # your own CrewAI LLM
    skills=decimalai.crewai.skills(),      # the installed folder
    tools=decimalai.crewai.skill_tools(),  # reads a skill's files, runs its scripts
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

No DecimalAI package. CrewAI's own loader reads the folder, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                         # your own CrewAI LLM
    skills=[Path(".agents/skills")],
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

CrewAI docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[adk]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

from decimalai.adk import DecimalAIPlugin, agent_config
from google.adk.agents import LlmAgent
from google.adk.apps import App

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = LlmAgent(
    name="support",          # ADK needs a Python identifier
    model="gemini-3.5-flash",
    instruction=support.instruction,  # your prompt, then the skill list
    tools=[*support.tools],           # your tools, then load_skill
)
# DecimalAIPlugin makes each invocation of runners built from this App one trace.
app = App(name="support", root_agent=agent, plugins=[DecimalAIPlugin()])

Prefer ADK's own loader? SkillToolset(skills=decimalai.adk.skills(), ...) reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

Install

pip install "decimalai[adk]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

from decimalai.adk import DecimalAIPlugin, agent_config
from google.adk.agents import LlmAgent
from google.adk.apps import App

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = LlmAgent(
    name="support",          # ADK needs a Python identifier
    model="gemini-3.5-flash",
    instruction=support.instruction,  # your prompt, then the skill list
    tools=[*support.tools],           # your tools, then load_skill
)
# DecimalAIPlugin makes each invocation of runners built from this App one trace.
app = App(name="support", root_agent=agent, plugins=[DecimalAIPlugin()])

Prefer ADK's own loader? SkillToolset(skills=decimalai.adk.skills(), ...) reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from google.adk.agents import LlmAgent

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = LlmAgent(
    name="support",
    model="gemini-3.5-flash",
    instruction=lambda context: f"You are a helpful agent.\n\n{skill}",
)

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Google ADK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[pydantic-ai]"
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

from pydantic_ai import Agent
from decimalai.pydantic_ai import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    "openai:gpt-4o-mini",
    name=support.name,
    tools=[],                 # your tools; the capability adds load_skill
    capabilities=[support],   # your prompt, the skill list and load_skill
)
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

On the Pydantic AI harness, decimalai.pydantic_ai_harness.skills(project) uses its own loader instead.

Install

pip install "decimalai[pydantic-ai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

from pydantic_ai import Agent
from decimalai.pydantic_ai import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    "openai:gpt-4o-mini",
    name=support.name,
    tools=[],                 # your tools; the capability adds load_skill
    capabilities=[support],   # your prompt, the skill list and load_skill
)
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

On the Pydantic AI harness, decimalai.pydantic_ai_harness.skills(project) uses its own loader instead.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from pydantic_ai import Agent

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = Agent("openai:gpt-4o-mini", instructions=f"You are a helpful agent.\n\n{skill}")
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Pydantic AI docs

Deep Agents' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" deepagents
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

import decimalai.deepagents
from deepagents import create_deep_agent

agent = create_deep_agent(model=model, **decimalai.deepagents.skills())
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

The default backend reads files but has no shell. To run a skill's scripts, pass skills(backend=LocalShellBackend(...)).

The same install from Python, for a notebook or a setup script. Deep Agents' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" deepagents

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

import decimalai.deepagents
from deepagents import create_deep_agent

agent = create_deep_agent(model=model, **decimalai.deepagents.skills())
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

The default backend reads files but has no shell. To run a skill's scripts, pass skills(backend=LocalShellBackend(...)).

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from deepagents import create_deep_agent

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = create_deep_agent(model=model, system_prompt=f"You are a helpful agent.\n\n{skill}")
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Deep Agents docs

Its own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" agent-framework
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

import asyncio

from agent_framework import Agent
from decimalai.agent_framework import skills

agent = Agent(client=client, instructions="You are a helpful agent.", context_providers=[skills()])
asyncio.run(agent.run("Hello!"))

It caches skills by default, so restart the process after decimalai skills update.

The same install from Python, for a notebook or a setup script. Its own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" agent-framework

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

import asyncio

from agent_framework import Agent
from decimalai.agent_framework import skills

agent = Agent(client=client, instructions="You are a helpful agent.", context_providers=[skills()])
asyncio.run(agent.run("Hello!"))

It caches skills by default, so restart the process after decimalai skills update.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from agent_framework import Agent

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = Agent(client=client, instructions=f"You are a helpful agent.\n\n{skill}")
asyncio.run(agent.run("Hello!"))

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Microsoft Agent Framework docs

Strands' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" strands-agents
decimalai skills add skydoves-using-strong-skipping-correctly

Add it to your agent

from decimalai.strands import skills
from strands import Agent
from strands.vended_tools import file_editor, shell

agent = Agent(
    model=model,                # always pass one: the default is Amazon Bedrock
    system_prompt="You are a helpful agent.",
    tools=[file_editor, shell],
    plugins=[skills()],
)
agent("Hello!")

Strands loads an agent's skills once, so build the agent again after decimalai skills update.

The same install from Python, for a notebook or a setup script. Strands' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" strands-agents

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("skydoves-using-strong-skipping-correctly")
if result.exit_code:
    raise SystemExit(result.problems)

from decimalai.strands import skills
from strands import Agent
from strands.vended_tools import file_editor, shell

agent = Agent(
    model=model,                # always pass one: the default is Amazon Bedrock
    system_prompt="You are a helpful agent.",
    tools=[file_editor, shell],
    plugins=[skills()],
)
agent("Hello!")

Strands loads an agent's skills once, so build the agent again after decimalai skills update.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

Give it to your agent

from pathlib import Path
from strands import Agent

skill = Path(".agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md").read_text()
agent = Agent(model=model, system_prompt=f"You are a helpful agent.\n\n{skill}")
agent("Hello!")

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Strands docs

Adds every file the skill has, pinned so you can update it. Cursor, Copilot, Cline and Warp read this folder by themselves.

pip install "decimalai>=0.15"
decimalai skills add skydoves-using-strong-skipping-correctly

No Python or account. Cursor, Copilot, Cline and Warp read this folder by themselves.

Save it into your project

curl -fsSL --create-dirs -o .agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md https://app.decimal.ai/s/skydoves-using-strong-skipping-correctly/SKILL.md

Copy SKILL.md

Save it in your project, as

.agents/skills/skydoves-using-strong-skipping-correctly/SKILL.md

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Cursor and other editors docs

Fetch the file and give its text to your model as instructions. No account or package needed.

Get the file

curl -fsSL https://app.decimal.ai/s/skydoves-using-strong-skipping-correctly/SKILL.md

Read the summary as JSON

curl -fsSL https://app.decimal.ai/s/skydoves-using-strong-skipping-correctly.json

This gets SKILL.md only. The skill's 1 other file is not included.

Other ways to add it

Other setups docs

You don't need a DecimalAI account.

Version: v1, tested 8/19/2026
Source: skydoves/compose-performance-skills ↗
License: Apache-2.0

Manage it in DecimalAI

Saved prompts and skills, pinned versions, and traces of this skill in your runs. DecimalAI does not run your agent.

Claude Code reads files from your project, so DecimalAI config does not apply there.

On Claude Agent SDK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On OpenAI Agents: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On OpenAI SDK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Anthropic: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Gemini API: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On LangChain: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On CrewAI: set DECIMAL_API_KEY and add decimalai.init(crewai=True) to record runs and skill use. The skill version stays in your repo.

On Google ADK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Pydantic AI: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Deep Agents: set DECIMAL_API_KEY and add decimalai.init(langchain=True) to record runs and skill use. The skill version stays in your repo.

On Microsoft Agent Framework: set DECIMAL_API_KEY and add decimalai.init(otel=True, agent_name="support") to record runs and skill use. The skill version stays in your repo.

On Strands: set DECIMAL_API_KEY and add decimalai.init(otel=True, agent_name="support") to record runs and skill use. The skill version stays in your repo.

Editors read files from your project, so DecimalAI config does not apply there.

Anywhere else: record a use with decimalai.log_skill_activation() inside a trace.

Use this skill

SKILL.md

Using Strong Skipping Correctly — make every restartable composable skippable, intentionally

Strong Skipping Mode is the Compose compiler behavior that became the default with the Kotlin Compose compiler plugin shipped in Kotlin 2.0.20. It changes two things at once: every restartable composable becomes skippable regardless of param stability (unstable params are compared with ===, stable params with equals), and every capturing lambda literal written inside a @Composable function is automatically wrapped in remember(captures) { ... }. Both behaviors are on by default; neither requires opt-in flags on Kotlin 2.0.20+. Lambdas with no captures are already compiler-emitted singletons and are NOT wrapped — only capturing lambdas need (and get) the auto-remember treatment.

That sounds like "stability no longer matters", and for a lot of UI code it almost is. But the change leaves three sharp edges: stable types still need correct equals semantics so the === fallback does not invalidate every recompose, lambdas in non-@Composable scopes (LazyListScope.items { }, Modifier.pointerInput { }, plain object expressions) are not auto-memoized, and a few intentional cases need explicit opt-out via @DontMemoize or @NonSkippableComposable. This skill walks the verification, audit, and escape-hatch decisions for working in a strong-skipping world.

When to use this skill

The developer asks "do I still need @Stable?" or "does this composable skip now?".
A composable still recomposes on every parent tick even though the developer "thought strong skipping fixed that".
A lambda allocation question comes up — "is this onClick = { vm.add(it) } allocating per recompose?".
Migrating a project from Kotlin 1.9.x / Compose Compiler 1.5.x to Kotlin 2.0.20+ and reconciling reports that now show restartable skippable everywhere.
The developer encounters @DontMemoize or @NonSkippableComposable in code or compiler output and asks what they do.
The user mentions "strong skipping", "auto-remember", "@DontMemoize", "@NonSkippableComposable", or "skippable everywhere".

When NOT to use this skill

The conceptual question is "what makes a type stable in the first place?" — use ../../stability/understanding-stability-inference/SKILL.md.
Stability is regressing silently and the team needs a CI gate — use ../../stability/enforcing-stability-in-ci/SKILL.md.
Even with strong skipping on, type-level stability still matters for diagnostic clarity and for equals-based skipping to work — use ../../stability/diagnosing-compose-stability/SKILL.md to fix the underlying types.
The recomposition is happening but the developer cannot see it — use ../debugging-recompositions/SKILL.md to instrument first.

Prerequisites

Kotlin 2.0.20+ with the Compose compiler Gradle plugin applied via org.jetbrains.kotlin.plugin.compose. Strong Skipping is on by default. To turn it off on Kotlin 2.0.20+, use composeCompiler { featureFlags.add(ComposeFeatureFlag.StrongSkipping.disabled()) } (the legacy enableStrongSkippingMode = true property is @Deprecated("Use the featureFlags option instead") and still works on older Kotlin).
Compose Compiler reports turned on for the module under audit (see ../../stability/diagnosing-compose-stability/SKILL.md).
Builds running in release when reading reports — debug adds Live Literals which distort skip behavior (see ../../measurement/testing-compose-in-release-mode/SKILL.md).
Familiarity with the per-composable line format in *-composables.txt (restartable skippable fun Foo(...)).

Workflow

1. Verify strong skipping is actually active

Before claiming any behavior is "because of strong skipping", confirm the toolchain is on a version where it is the default, or the legacy flag is set.

./gradlew :app:dependencies --configuration kotlinCompilerPluginClasspathRelease | grep -i compose

Expect the Compose compiler plugin to come from Kotlin 2.0.20 or later. In build.gradle.kts:

plugins {
    alias(libs.plugins.kotlin.android)        // Kotlin 2.0.20+
    alias(libs.plugins.kotlin.compose)        // org.jetbrains.kotlin.plugin.compose
}

To disable strong skipping on Kotlin 2.0.20+ (the modern feature-flag DSL):

import androidx.compose.compiler.plugins.kotlin.ComposeFeatureFlag

composeCompiler {
    featureFlags.add(ComposeFeatureFlag.StrongSkipping.disabled())
}

The legacy property composeCompiler { enableStrongSkippingMode = true } is @Deprecated("Use the featureFlags option instead") in modern Compose Compiler Gradle plugin releases. It still works (and is still the only switch on older Kotlin 1.9.x / 2.0.0–2.0.10), but new code should prefer featureFlags.add(ComposeFeatureFlag.StrongSkipping.disabled()) when opting out.

Then regenerate Compose Compiler reports and grep for a known previously-unstable signature:

./gradlew :app:assembleRelease
grep -A1 "restartable" app/build/compose_compiler/app-release-composables.txt | head -40

Under strong skipping, every restartable composable should print restartable skippable, including ones that take List<Foo> or other unstable parameters. If a composable is still listed restartable without skippable, it is using @NonSkippableComposable (intentional) or the mode has been turned off (e.g. someone added featureFlags.add(ComposeFeatureFlag.StrongSkipping.disabled()) or set the legacy enableStrongSkippingMode = false).

2. Audit lambda capture sites inside @Composable scopes

Strong skipping wraps every capturing lambda literal written inside a @Composable function in remember(captures) { ... }. The captured-values list is the compiler-generated remember key. Capture-less lambdas (e.g. { vm.refresh() } where vm is a stable file-scope reference, or { /* no-op */ }) are already compiler-emitted singletons and are NOT auto-remembered — there is nothing to key on. Three consequences:

A capture-less lambda is a single shared instance across all recompositions of the surrounding function (singleton emitted by the compiler).
A lambda that captures only stable values is memoized once via remember(captures) and re-used across recompositions while the captures compare equal.
A lambda that captures an unstable value (e.g. a List<Foo> literal not hoisted) re-captures on every recompose because the unstable capture compares with ===, and a fresh literal is a new reference.

Inspect the suspect call site. The mental model: "what does this lambda close over, and is each capture stable by equals and identity-stable across recompositions?".

// RIGHT — captures stable id + stable callback; auto-remembered by strong skipping
@Composable
fun Row(snack: Snack, onClick: (Long) -> Unit) {
    Button(onClick = { onClick(snack.id) }) { Text(snack.name) }
}

// WRONG — captures a freshly allocated list literal; new reference each recompose
@Composable
fun Header(title: String) {
    Toolbar(actions = listOf(Action.Share, Action.Save)) { /* ... */ }
}
// WRONG because: `listOf(...)` allocates a new List every recompose. Even with strong
// skipping the unstable param compares with ===, fails, and Toolbar recomposes every time.

// RIGHT — hoist the unstable literal so its identity is stable
@Composable
fun Header(title: String) {
    val actions = remember { listOf(Action.Share, Action.Save) }
    Toolbar(actions = actions) { /* ... */ }
}

3. Identify the gaps where strong skipping does NOT memoize

Auto-remember only applies to lambdas inside a @Composable function context. The following are not @Composable scopes; lambdas inside them are not auto-remembered and need a manual remember(key) if allocation matters:

LazyListScope.items { ... }, LazyGridScope.items { ... }, LazyVerticalGrid.items { ... }, etc. — the items block is a regular Kotlin DSL builder.
Modifier.pointerInput { ... }, Modifier.draggable { ... }, Modifier.scrollable { ... } — non-@Composable lambdas.
Object expressions and SAM conversions (e.g. object : DefaultLifecycleObserver { }).
Modifier.drawBehind { ... }, Modifier.drawWithCache { ... } — the lambda parameters are not @Composable.
Coroutine builders inside non-@Composable scopes.

For these, hoist a stable lambda yourself, keyed on whatever may change:

// WRONG — onClick allocated per item per recompose; strong skipping does not reach here
LazyColumn {
    items(snacks, key = { it.id }) { snack ->
        SnackRow(snack, onClick = { vm.select(snack.id) })
    }
}
// WRONG because: `items { snack -> ... }` is LazyListScope, not a @Composable function.
// Strong skipping does not auto-remember the onClick lambda — a new instance allocates per row,
// per recompose, and SnackRow's `===` comparison on the lambda param fails every time.

// RIGHT — hoist a stable handler keyed on the stable callback
LazyColumn {
    items(snacks, key = { it.id }) { snack ->
        val onClick = remember(snack.id) { { vm.select(snack.id) } }
        SnackRow(snack, onClick = onClick)
    }
}

// RIGHT — even better: pass the id-aware callback once
LazyColumn {
    items(snacks, key = { it.id }) { snack ->
        SnackRow(snack, onClick = vm::select)   // method reference is stable
    }
}

4. Decide on escape hatches

Strong skipping is opinionated. A handful of cases legitimately need the opposite behavior. The full reference table is in references/escape-hatches.md; the high-level decision tree:

@DontMemoize at a lambda expression site (the annotation targets EXPRESSION only, not TYPE) — the lambda must NOT be auto-remembered. Use when the lambda intentionally captures a fresh value per call (e.g. a logger that should record the latest counter, not the one captured by the first composition). Apply it at the call site: Button(onClick = @DontMemoize { ... }), never at a parameter type position like (@DontMemoize () -> Unit).
@NonSkippableComposable on a function — opt the function out of skipping entirely. Use for side-effect-only composables (a logger, a debug overlay) that must run on every recompose regardless of param equality.
@NonRestartableComposable on a function — drop the restart scope entirely. Use only for very small composables where the restart bookkeeping costs more than the function body. Almost never needed in app code; Compose UI uses it internally.
@ReadOnlyComposable on a function — declare the function only reads composition values and produces no nodes, so it can run in initialization contexts (e.g. computing a value to pass into another composable). Compose's MaterialTheme.colorScheme getters use it.

// RIGHT — logger composable that must always emit
@NonSkippableComposable
@Composable
fun RecompositionLogger(label: String) {
    SideEffect { Log.d("Recomp", "tick: $label @ ${System.currentTimeMillis()}") }
}

// RIGHT — annotate the lambda expression that intentionally captures fresh per-call state.
// @DontMemoize is @Target(EXPRESSION) — apply it on the lambda literal, NOT on the type.
@Composable
fun TraceClicks(counter: Int, content: @Composable (() -> Unit) -> Unit) {
    content(@DontMemoize { Log.d("Click", "counter at click time = $counter") })
}

5. Prove behavior with @TraceRecomposition

Reading reports tells the story up to the function boundary. To confirm a specific composable is or is not being skipped at runtime, instrument with @TraceRecomposition from compose-stability-analyzer. Cross-link ../../measurement/tracing-recompositions-at-runtime/SKILL.md for the full setup. The two-line summary:

@TraceRecomposition(traceStates = true)
@Composable
fun SnackRow(snack: Snack, onClick: (Long) -> Unit) { /* ... */ }

Logcat will print one line per recomposition with the changed parameters, which directly answers "did strong skipping skip this call or not?".

Patterns

Pattern: stability still matters even under strong skipping

Strong skipping does not eliminate the need for @Immutable / @Stable. It changes the comparison from equals to === for unstable params. === is reference equality. If a stable type is needed in a hot path, mark it @Immutable so equals is used and structurally-equal instances skip.

// WRONG
data class Filter(val tags: List<String>, val sort: SortOrder)

@Composable
fun FilterBar(filter: Filter) { /* ... */ }
// WRONG because: Filter is unstable (List<String> is unstable). Strong skipping makes
// FilterBar skippable, but only when `prevFilter === newFilter`. A copy() with the same
// content fails identity, recomposes the bar, every time.

// RIGHT
@Immutable
data class Filter(val tags: ImmutableList<String>, val sort: SortOrder)

@Composable
fun FilterBar(filter: Filter) { /* ... */ }
// RIGHT because: @Immutable + ImmutableList makes Filter stable. Strong skipping uses
// equals(), and structurally-equal copies skip.

Pattern: literal collection in a hot composable param

// WRONG
@Composable
fun Toolbar(title: String) {
    ActionRow(actions = listOf(Action.Share, Action.Save))
}
// WRONG because: a fresh List allocates per recompose. Strong skipping makes ActionRow
// skippable, but the List param fails === every time and the body re-runs.

// RIGHT
private val ToolbarActions = persistentListOf(Action.Share, Action.Save)

@Composable
fun Toolbar(title: String) {
    ActionRow(actions = ToolbarActions)
}

Pattern: lambda inside LazyListScope.items

See workflow step 3. The short version: items { } is not @Composable, strong skipping does not reach inside it, hoist a stable callback or use a method reference.

Pattern: `@DontMemoize` for intentionally-fresh lambdas

// WRONG — auto-remembered lambda captures `counter` from the FIRST composition
@Composable
fun LoggingButton(counter: Int) {
    Button(onClick = { Log.d("Btn", "counter = $counter") }) { Text("hit") }
}
// WRONG because: under strong skipping the lambda is auto-remembered, but `counter` is a
// stable Int capture, so the remember key is the int value. This is actually fine here.
// Where it bites is when the developer EXPECTS each invocation to capture fresh state but
// the lambda is being passed across a memoization boundary that hides the capture.

// RIGHT — explicit opt-out where freshness is the contract.
// @DontMemoize is @Target(EXPRESSION) — apply on the lambda literal, never on the type.
@Composable
fun TraceContent(latest: Int, body: @Composable (() -> Unit) -> Unit) {
    body(@DontMemoize { sendTelemetry("snapshot=$latest") })
}

@DontMemoize is rarely needed in app code. Reach for it only after measuring; the default behavior is correct for almost all UI.

Pattern: `@NonSkippableComposable` for side-effect-only composables

// RIGHT
@NonSkippableComposable
@Composable
fun FrameTicker(onTick: () -> Unit) {
    SideEffect { onTick() }
}

The annotation says "this composable has work to do every recomposition; do not let the skip guard short-circuit it". Use sparingly — most composables that "must run every time" are actually misplaced state reads; fix the read site first.

Pattern: pair with CI to prevent strong-skipping regressions

Strong skipping makes nearly every composable skippable in the report, which makes the report less useful for spotting regressions by eye. Combine with ../../stability/enforcing-stability-in-ci/SKILL.md so the baseline diff catches a composable that flips from skippable to not.

Mandatory rules

MUST verify the Kotlin / Compose compiler plugin version before claiming strong skipping is active. Print ./gradlew :app:dependencies and confirm Kotlin 2.0.20+ (default ON). On older Kotlin, confirm the legacy enableStrongSkippingMode = true flag in composeCompiler { }. If the project explicitly disables Strong Skipping via featureFlags.add(ComposeFeatureFlag.StrongSkipping.disabled()), treat the rest of this skill as not applicable.
MUST NOT declare a composable "skippable now thanks to strong skipping" without checking the actual *-composables.txt report. Annotations like @NonSkippableComposable and lambdas inside LazyListScope.items { } remain footguns the report makes visible.
MUST prefer @Immutable / @Stable over relying on reference equality for unstable types passed in hot paths. === is a poor substitute for structural equals when the producer of the value (copy(), map { }, builder DSL) does not preserve identity.
MUST NOT use @DontMemoize or @NonSkippableComposable without a one-line code comment explaining why. These are advanced opt-outs; the next reader will assume they are mistakes if the rationale is not co-located.
MUST NOT treat strong skipping as a substitute for hoisting unstable literals. A listOf(...) allocated in a composable param fails === regardless of strong skipping. Hoist into remember { ... } or a top-level persistentListOf(...).
PREFERRED: combine this skill with ../../stability/enforcing-stability-in-ci/SKILL.md so a composable regressing from skippable to not skippable (e.g. someone adds @NonSkippableComposable without justification) fails CI.
PREFERRED: when a developer asks "is my lambda allocated per recompose?", confirm the call site is inside a @Composable function before answering yes/no. The answer flips entirely at the @Composable boundary.

Skippability is a diagnostic, not a KPI: under strong skipping every composable is "skippable" in the report, but only the ones whose params actually compare equal will skip at runtime. The interesting question is no longer "is it skippable?" but "do its params compare equal across the recomposition the developer cares about?".

Verification

Confirmed Kotlin / Compose compiler version is 2.0.20+ (default ON) and that no featureFlags.add(ComposeFeatureFlag.StrongSkipping.disabled()) opt-out is in effect — or, on older Kotlin, that the legacy enableStrongSkippingMode = true flag is set.
*-composables.txt shows restartable skippable on previously-unstable-param composables.
Audited the suspect call site for unstable literal captures (listOf, mapOf, ad-hoc lambdas in non-@Composable scopes).
Any @DontMemoize / @NonSkippableComposable / @NonRestartableComposable / @ReadOnlyComposable in the changed code has a co-located comment explaining why.
Runtime confirmed via @TraceRecomposition (see ../../measurement/tracing-recompositions-at-runtime/SKILL.md) that the composable skips when its inputs are equal.
Stability baseline (see ../../stability/enforcing-stability-in-ci/SKILL.md) covers the changed module so a future regression in skip behavior fails CI.

References

Strong Skipping (Android Developers) — https://developer.android.com/develop/ui/compose/performance/stability/strongskipping
Ben Trengrove, "Jetpack Compose Strong Skipping Mode explained" — https://medium.com/androiddevelopers/jetpack-compose-strong-skipping-mode-explained-cbdb2aa4b900
Ben Trengrove, "New ways of optimizing stability in Jetpack Compose" — https://medium.com/androiddevelopers/new-ways-of-optimizing-stability-in-jetpack-compose-038106c283cc
Compose Compiler release notes — https://developer.android.com/jetpack/androidx/releases/compose-compiler
Compose stability overview — https://developer.android.com/develop/ui/compose/performance/stability
Chris Banes, "Composable metrics" — https://chrisbanes.me/posts/composable-metrics/
skydoves, "Optimize App Performance by Mastering Stability" — https://medium.com/proandroiddev/optimize-app-performance-by-mastering-stability-in-jetpack-compose-69f40a8c785d
skydoves/compose-stability-analyzer — https://github.com/skydoves/compose-stability-analyzer
references/escape-hatches.md — full reference table of @NonRestartableComposable, @NonSkippableComposable, @DontMemoize, @ReadOnlyComposable with code samples and decision matrix.

For diagnosing why a composable is unstable in the first place, see ../../stability/diagnosing-compose-stability/SKILL.md. For instrumenting a specific composable to see if it actually skips at runtime, see ../debugging-recompositions/SKILL.md and ../../measurement/tracing-recompositions-at-runtime/SKILL.md. For preventing a strong-skipping regression in CI, see ../../stability/enforcing-stability-in-ci/SKILL.md.

About this skill

Automatic

Frontend / UI pack

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

skydoves / compose-performance-skills

recomposition/using-strong-skipping-correctly/SKILL.md

View on GitHub ↗

License Apache-2.0. See LICENSE for complete terms.Imported 8/3/2026Version v1Bundled 1 file

Illustrative Want to see what skydoves/using-strong-skipping-correctly would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Other skills for the same job

Published skills whose descriptions closely match this one, with their headline benchmark lift.

Related skills

Skills that do similar work, with their headline benchmark lift.

From the same source and nearby in the catalogue.

Ready to supercharge your AI agents?

SkillScore

Safety

What it changes

Track record

Use this skill

Using Strong Skipping Correctly — make every restartable composable skippable, intentionally

When to use this skill

When NOT to use this skill

Prerequisites

Workflow

1. Verify strong skipping is actually active

2. Audit lambda capture sites inside @Composable scopes

3. Identify the gaps where strong skipping does NOT memoize

4. Decide on escape hatches

5. Prove behavior with @TraceRecomposition

Patterns

Pattern: stability still matters even under strong skipping

Pattern: literal collection in a hot composable param

Pattern: lambda inside LazyListScope.items

Pattern: @DontMemoize for intentionally-fresh lambdas

Pattern: @NonSkippableComposable for side-effect-only composables

Pattern: pair with CI to prevent strong-skipping regressions

Mandatory rules

Verification

References

SkillSafety

Source

Other skills for the same job

Related skills

More in the registry

Pattern: `@DontMemoize` for intentionally-fresh lambdas

Pattern: `@NonSkippableComposable` for side-effect-only composables