Ready to supercharge your AI agents?

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

Get Started Free →

Imaging Data Commons · k-dense-ai | DecimalAI

Scienceby k-dense-ai

Imaging Data Commons

Query and download public cancer imaging data from NCI Imaging Data Commons using idc-index.

Never made a case worseMaintained

SkillScore

SkillScore100/100

Test results 100

Safety

3/3 safety checks passed

Code scan
Intent
Content

What it changes

26%100%of test cases passed +74 pts

Without skill6 of 23

With skill23 of 23

17 fixed6 passed either way0 still failing0 broke

Measured on gemini-3.6-flash8/17/2026v2 · latest is v322 comparable

This benchmark measured an earlier version of the skill.

Track record

Agent runs: 0; All time

See one task, with and without the skill

One test taskBefore attempting to install or run Python code with `idc-index`, what script shipped in the IDC workflow should be executed to verify version compatibility and get the exact recommended installation command for the current interpreter?

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 k-dense-ai-imaging-data-commons
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("k-dense-ai-imaging-data-commons")
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 k-dense-ai-imaging-data-commons

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/k-dense-ai-imaging-data-commons/SKILL.md https://app.decimal.ai/s/k-dense-ai-imaging-data-commons/SKILL.md

Copy SKILL.md

Save it in your project, as

.claude/skills/k-dense-ai-imaging-data-commons/SKILL.md

This gets SKILL.md only. The skill's 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

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

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
from agents import Agent, Runner

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
from openai import OpenAI

client = OpenAI()
skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
import anthropic

client = anthropic.Anthropic()
skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/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/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
from langchain.agents import create_agent

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

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

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
from pydantic_ai import Agent

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
from deepagents import create_deep_agent

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from agent_framework import Agent

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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("k-dense-ai-imaging-data-commons")
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/k-dense-ai-imaging-data-commons/SKILL.md

Give it to your agent

from pathlib import Path
from strands import Agent

skill = Path(".agents/skills/k-dense-ai-imaging-data-commons/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 14 other files are 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 k-dense-ai-imaging-data-commons

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/k-dense-ai-imaging-data-commons/SKILL.md https://app.decimal.ai/s/k-dense-ai-imaging-data-commons/SKILL.md

Copy SKILL.md

Save it in your project, as

.agents/skills/k-dense-ai-imaging-data-commons/SKILL.md

This gets SKILL.md only. The skill's 14 other files are 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/k-dense-ai-imaging-data-commons/SKILL.md

Read the summary as JSON

curl -fsSL https://app.decimal.ai/s/k-dense-ai-imaging-data-commons.json

This gets SKILL.md only. The skill's 14 other files are not included.

Other ways to add it

Other setups docs

You don't need a DecimalAI account.

Version: v3 (v2 tested 8/17/2026)
Source: k-dense-ai/scientific-agent-skills ↗
License: MIT

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

Imaging Data Commons

Overview

Query and download public cancer imaging data from the National Cancer Institute Imaging Data Commons (IDC). No authentication required for data access.

Expected network access: IDC metadata is reachable three ways — a bundled local DuckDB index (offline after installation; additional indices are fetched from GitHub and clinical tables from S3), or the hosted IDC service over MCP or REST (api.imaging.datacommons.cancer.gov, no authentication). File downloads use public GCS (storage.googleapis.com) and AWS S3 (s3.amazonaws.com) — no authentication required. DICOMweb access uses either the public IDC proxy (proxy.imaging.datacommons.cancer.gov, no auth) or the Google Cloud Healthcare API (healthcare.googleapis.com, requires GCP authentication). Optional BigQuery queries (bigquery.googleapis.com) also require GCP authentication. Citation resolution contacts DOI services. Public IDC routes require no credentials; optional Google clients use Application Default Credentials.

Reviewed 2026-09-30: idc-index 0.12.5, idc-index-data 24.2.2, IDC v24; hosted API 3.0.0b3. Recheck at use time.

Choose the access path first. There is no single default: the cheapest correct path depends on the session and the task.

Session already has the IDC MCP server? Route discovery and metadata there — see IDC MCP Server.
Otherwise, is idc-index installed and current? Run python scripts/check_version.py. If it passes, use idc-index for everything.
Not installed, and the task is read-only metadata — counts, attribute values, collection lookups, SQL under 10 000 rows, licenses, citations, viewer URLs? Use the REST API over curl; do not install anything. Installing costs ~77 MB of packaged index data plus pandas, pyarrow, and duckdb, which a metadata question does not need. See Data Access Options.
Not installed, and the task needs more than metadata — downloading files, pandas or plotting, pydicom/SimpleITK, pathology tiling, results past 10 000 rows, or a version-pinned script the user re-runs? Install idc-index: check_version.py exits non-zero and prints the exact install command for the running interpreter. Prefer a virtual environment, then restart Python.

idc-index (GitHub) is still the most capable Python path, with query and download helpers in one client. check_version.py never installs anything itself — it also flags a newer idc-index or skill release when one exists.

Setup for the idc-index path: use the intended interpreter for scripts/check_version.py and confirm its meets pinned minimum message before continuing. A launch failure is not a pass.

from idc_index import IDCClient
client = IDCClient()

# Verify IDC data version (should be "v24")
print(f"IDC data version: {client.get_idc_version()}")

Download and image-processing examples are illustrative unless noted; see the reference review notes for verification scope.

Core workflow: query metadata with client.sql_query() → download with client.download_from_selection() → visualize with client.get_viewer_URL(). Python examples below assume this client; Data Access Options has the REST equivalents. For current data scale, run the summary query in references/sql_patterns.md or GET /v3/stats.

IDC MCP Server

IDC operates a hosted MCP server at https://api.imaging.datacommons.cancer.gov/mcp (streamable HTTP, no authentication). Where it is available it complements — it does not replace — the idc-index workflow below.

Identify it by the MCP resource idc://guide, or by three or more of the tool names build_cohort, get_cohort_urls, list_analysis_results, and get_idc_version. Generic names such as run_sql are not evidence on their own. If identification is ambiguous, use idc-index.

If this session has the server, treat it as authoritative for discovery and metadata — IDC version, counts, attribute values, cohort building, metadata SQL — and follow the server's own instructions rather than re-deriving them from this file. Its data version is whatever the server reports: call get_idc_version instead of relying on the version pinned in this file.

Return here for what the server does not do: downloading files, local pandas/notebook analysis, DICOMweb, BigQuery, digital pathology tiling, and reproducible scripts. Hand off by passing SeriesInstanceUIDs from the server to client.download_from_selection(...), and run scripts/check_version.py at that point.

If it is not available, the identical service is reachable with no configuration as a REST API at https://api.imaging.datacommons.cancer.gov/v3 — use it for read-only metadata rather than installing idc-index, per the routing gate in Overview. Suggest connecting the MCP server at most once, only for repeated interactive discovery, and never change the user's configuration yourself.

See references/mcp_guide.md for the tool inventory, handoff patterns, and per-host notes.

When to Use This Skill

Finding publicly available radiology (CT, MR, PET) or pathology (slide microscopy) images
Selecting image subsets by cancer type, modality, anatomical site, or other metadata
Downloading DICOM data from IDC
Checking data licenses before use in research or commercial applications
Visualizing medical images in a browser without local DICOM viewer software

Quick Navigation

Inline below: the MCP/REST routing rules, the IDC data model, the index tables and how they join, the core API patterns (query, download, visualize, license, cite), best practices, and troubleshooting.

Reference Guides (load on demand):

Guide	When to Load
`index_tables_guide.md`	Complex JOINs, schema discovery, DataFrame access
`use_cases.md`	End-to-end workflows: training datasets, batch downloads, DICOM reading with pydicom/SimpleITK, pipeline integration
`sql_patterns.md`	Quick SQL patterns for filter discovery, annotations, size estimation
`clinical_data_guide.md`	Clinical/tabular data, imaging+clinical joins, value mapping
`licensing_and_citation.md`	Commercial-use questions, mixed-license cohorts, citation formats
`cloud_storage_guide.md`	Direct S3/GCS access, versioning, UUID mapping
`dicomweb_guide.md`	DICOMweb endpoints, PACS integration
`digital_pathology_guide.md`	Slide microscopy (SM), annotations (ANN), pathology workflows
`bigquery_guide.md`	Full DICOM metadata, private elements (requires GCP)
`cli_guide.md`	Command-line tools (`idc download`, manifest files)
`parquet_access_guide.md`	Direct Parquet queries via GCS (no idc-index install needed)
`mcp_guide.md`	Hosted IDC MCP server: tool inventory, identification, handoff to `idc-index`
`rest_api_guide.md`	Hosted IDC REST API: endpoints, filter syntax, SQL over HTTP, manifests

IDC Data Model

IDC adds two grouping levels above the standard DICOM hierarchy (Patient → Study → Series → Instance):

collection_id: Groups patients by disease, modality, or research focus (e.g., tcga_luad, nlst). Treat (collection_id, PatientID) as the patient key; do not assume PatientID is globally unique.
analysis_result_id: Identifies derived objects (segmentations, annotations, radiomics features) across one or more original collections. Use it to find AI-generated or expert annotations, while collection_id finds original imaging data (which may itself include deposited annotations).

Key identifiers for queries:

Identifier	Scope	Use for
`collection_id`	Dataset grouping	Filtering by project/study
`PatientID`	Patient	Grouping images by patient
`StudyInstanceUID`	DICOM study	Grouping of related series, visualization
`SeriesInstanceUID`	DICOM series	Grouping of related series, visualization

Index Tables

The idc-index package provides multiple metadata index tables, accessible via SQL or as pandas DataFrames. The REST API exposes the same tables through GET /tables and POST /sql.

Important: client.indices_overview is the authoritative source for current table descriptions, available columns, and their types — query it when writing SQL or exploring data structure. It also answers "which table contains column X"; see references/index_tables_guide.md for that search pattern and full schema discovery.

Available Tables

Always call client.fetch_index("table_name") before querying any index table — it is safe and idempotent for all tables, including those loaded automatically at startup.

Family	Tables	Granularity
Core	`index` (primary metadata for all current data), `collections_index`, `analysis_results_index`	series / collection / analysis result
Modality acquisition parameters	`ct_index`, `mr_index`, `pt_index`, `contrast_index`	1 row = 1 series of that modality
Derived objects	`seg_index`, `rtstruct_index`, `ann_index`, `ann_group_index`	1 row = 1 series (or annotation group)
Microscopy	`sm_index`, `sm_instance_index`	1 row = 1 SM series / instance
Geometry, clinical, history	`volume_geometry_index`, `clinical_index`, `version_metadata_index`, `prior_versions_index`	see guide

references/index_tables_guide.md has the full inventory with each table's columns and contents — load it when you need to know what a specialized table actually holds.

prior_versions_index contains historical series versions, including revised versions whose DICOM SeriesInstanceUID still occurs in index. Pin crdc_series_uuid for historical content. For "what's new" in the current release use series_init_idc_version / series_revised_idc_version in the main index table, which are not equivalent to this table's min_idc_version / max_idc_version.

Joining Tables

SeriesInstanceUID is the universal join key for all series-level specialized tables: sm_index, sm_instance_index, seg_index, ann_index, ann_group_index, contrast_index, volume_geometry_index, rtstruct_index, ct_index, mr_index, pt_index. Always join these to index on SeriesInstanceUID. The exceptions below use different column names.

Join Column	Tables	Use Case
`collection_id`	index, prior_versions_index, collections_index, clinical_index	Link series to collection metadata or clinical data
`analysis_result_id`	index, analysis_results_index	Link series to analysis result metadata (annotations, segmentations)
`source_DOI`	index, analysis_results_index	Link by publication DOI
`segmented_SeriesInstanceUID`	seg_index → index	Link segmentation to its source image series (`seg_index.segmented_SeriesInstanceUID = index.SeriesInstanceUID`)
`referenced_SeriesInstanceUID`	ann_index → index, rtstruct_index → index	Link annotation or RTSTRUCT to its source image series

Note: subjects, updated, and description appear in multiple tables but have different meanings (counts vs identifiers, different update contexts). A UID-only join to prior_versions_index can match several historical revisions; use CRDC UUIDs to distinguish them.

For detailed join examples, schema discovery patterns, key columns reference, and DataFrame access, see references/index_tables_guide.md.

Clinical Data Access

Clinical (non-imaging) attributes — staging, demographics, therapy — live in per-collection tables. client.fetch_index("clinical_index") loads the dictionary mapping columns to collections; client.get_clinical_table(name) returns one table as a DataFrame.

See references/clinical_data_guide.md for the discovery workflow, coded-value mapping, and joining clinical data with imaging.

Data Access Options

Method	Auth	Best For	Reference
`idc-index`	No	Downloads, pandas analysis, unbounded queries — the most capable path	This document
IDC MCP server	No	Discovery, cohort building, metadata when the session already has it	`mcp_guide.md`
IDC REST API	No	Metadata with no install, from any language or shell — the default when `idc-index` is absent	`rest_api_guide.md`
Direct Parquet (GCS)	No	Version-pinned queries, or results past the REST row cap	`parquet_access_guide.md`
Cloud storage (S3/GCS)	No	Direct file access, bulk transfer, custom pipelines	`cloud_storage_guide.md`
DICOMweb via IDC proxy	No	Tool and PACS integration; daily quota, so testing and moderate use	`dicomweb_guide.md`
DICOMweb via Google Healthcare	Yes (GCP)	The same DICOMweb API at production volume, without the proxy quota	`dicomweb_guide.md`
SlicerIDCBrowser	No	3D visualization and analysis in 3D Slicer	https://github.com/ImagingDataCommons/SlicerIDCBrowser
BigQuery	Yes (GCP)	Full DICOM metadata, private elements, SR measurements — last resort	`bigquery_guide.md`

The IDC Portal (https://portal.imaging.datacommons.cancer.gov/) is interactive only — browser-based exploration, manual cohort selection, and download. Unlike every option above it has no programmatic interface, so point a user there to browse or click through data themselves; never use it as a step in a script or workflow.

REST API — the no-install metadata path

https://api.imaging.datacommons.cancer.gov/v3, no authentication: discovery, cohort counts and manifests, read-only SQL, clinical tables, viewer URLs, licenses, citations. It is the same service as the MCP server over plain HTTP, so it needs no configuration. It never moves image bytes — switch to idc-index to download, to get a DataFrame, or for results past 10 000 rows.

B=https://api.imaging.datacommons.cancer.gov/v3
curl -s $B/version   # idc_version, idc_index_data_version, api_version
curl -s $B/stats     # collections, patients, studies, series, instances, size_TB
curl -s "$B/attributes/Modality/values?limit=5"   # real filter values, with counts
curl -s $B/sql -H 'content-type: application/json' \
  -d '{"sql":"SELECT collection_id, COUNT(*) n FROM index GROUP BY 1 ORDER BY n DESC LIMIT 3"}'
curl -s $B/cohort/counts -H 'content-type: application/json' \
  -d '{"filters":{"terms":{"collection_id":["rider_pilot"]}}}'

The filter object always goes under filters — on cohort/counts, cohort/manifest, cohort/manifest.txt, licenses, and citations alike. A bare filter or an unrecognized key is a 422 naming the fix; an unfiltered series-enumerating request is a 400, not the whole archive. Filtered JSON responses echo filters_applied and warnings (inside counts for manifests) — read them, because they name any predicate the server dropped. A zero count with empty warnings therefore means the filter matched nothing, not that a value was miscased; miscasing produces a warning that says so.

POST /sql takes one read-only SELECT/WITH over the tables idc-index exposes plus clinical.<table>; max_rows defaults to 5 000, caps at 10 000, and truncated flags clipping. GET /attributes lists the 19 filterable attributes — clinical values, segmented anatomy, and acquisition parameters are not among them and need SQL. Request limits still apply. Use v3 only: V1 and V2 are superseded and scheduled for shutdown, so port any /v1/- or Modality_btw-style example a user brings rather than extending it.

Both sides build on idc-index-data, so compare the API's idc_index_data_version against local idc_index_data.__version__ before mixing them: the major is the IDC data release (24.x.y serves v24); minor/patch index builds may correct metadata. If the API is a whole release ahead, idc-index cannot download the extra series — mixed selections can omit unrecognized UIDs, while wholly unmatched selections raise — so either upgrade it (run scripts/check_version.py for the right command) or transfer directly from the bucket with s5cmd --no-sign-request.

See references/rest_api_guide.md for the endpoint reference, filter grounding, limits, and the manifest-based download flow.

Cloud storage organization

All DICOM files live in public buckets mirrored between AWS S3 and GCS, organized by CRDC UUIDs (not DICOM UIDs) to support versioning, as <crdc_series_uuid>/<crdc_instance_uuid>.dcm. Access is free (no egress fees) via AWS CLI, gsutil, or s5cmd with anonymous access; use the series_aws_url column for S3 URLs. Bucket names do not establish a license; query license_short_name for each selected series. See references/cloud_storage_guide.md for the full bucket list and UUID mapping.

DICOMweb access

IDC data is available via DICOMweb (Google Cloud Healthcare API) for PACS integration and DICOMweb-compatible tools: a public proxy (no auth, daily quota) for testing and moderate queries, or Google Healthcare (GCP auth) for production volumes. See references/dicomweb_guide.md.

Direct Parquet access

The idc-index metadata tables are also published as Parquet on a public GCS bucket (idc-index-data-artifacts), queryable with DuckDB or pandas. This needs DuckDB installed For ad-hoc metadata prefer REST /sql; choose Parquet for pinned versions or large results. A separate public S3 export also includes clinical tables and full BigQuery metadata. See references/parquet_access_guide.md.

Core Capabilities

The patterns below are the ones that go wrong when recalled from memory rather than checked. Worked examples for each area live in the reference guides named inline.

1. Discovery — enumerate values before filtering on them

Filtering on a guessed Modality or BodyPartExamined string is the most common cause of an empty result set. Enumerate first:

modalities = client.sql_query("""
    SELECT DISTINCT Modality, COUNT(*) as series_count
    FROM index
    GROUP BY Modality
    ORDER BY series_count DESC
""")
print(modalities)

The same pattern works for any filter column, optionally narrowed by another — BodyPartExamined within a Modality, Manufacturer, collection_id. On the REST path this grounding is a single call — GET /attributes/{attr}/values returns values with counts — and the cohort endpoints report a miscased value in warnings rather than as an empty result.

Two indices carry curated collection-level metadata the primary index does not, both requiring client.fetch_index(...) first: collections_index (cancer types, tumor locations, species, subject counts) and analysis_results_index (derived datasets — AI segmentations, expert annotations, radiomics — with their source collections and modalities).

Cancer type lives in collections_index.cancer_types, not in index — filtering by cancer type requires a join:

client.fetch_index("collections_index")
results = client.sql_query("""
    SELECT i.collection_id, i.PatientID, i.SeriesInstanceUID, i.Modality
    FROM index i
    JOIN collections_index c ON i.collection_id = c.collection_id
    WHERE c.cancer_types LIKE '%Breast%'
      AND i.Modality = 'MR'
    LIMIT 20
""")

client.sql_query() returns a pandas DataFrame. Confirm column names with client.get_index_schema('index') or client.indices_overview before writing a query rather than assuming them.

See references/sql_patterns.md for filter-value discovery, annotation and segmentation queries, size estimation, clinical linking, and version tracking ("what's new in vX" — use series_init_idc_version / series_revised_idc_version in index; historical objects use prior_versions_index).

2. Downloading DICOM files

The two download methods take their first two arguments in opposite order. This is the most common source of broken IDC code — check it rather than recalling it:

Method	First arg	Second arg	Use when
`download_from_selection`	`downloadDir` (required)	filter kwargs (optional)	Filtering by collection, patient, study, or series
`download_dicom_series`	`seriesInstanceUID` (required)	`downloadDir` (required)	Downloading specific series by UID only

download_from_selection takes filter keyword arguments, NOT a DataFrame. The name "from_selection" refers to filtering the IDC index by criteria — not to accepting a pandas DataFrame. To download query results, extract the UIDs into a list first:

# Step 1: Query for series UIDs
series_df = client.sql_query("""
    SELECT SeriesInstanceUID
    FROM index
    WHERE Modality = 'CT'
      AND BodyPartExamined = 'CHEST'
      AND collection_id = 'nlst'
    LIMIT 5
""")

# Step 2: Extract UIDs as a list from the DataFrame
uids = list(series_df['SeriesInstanceUID'].values)

# Step 3: Pass the list to download_from_selection (NOT the DataFrame itself)
client.download_from_selection(
    downloadDir="./data/lung_ct",
    seriesInstanceUID=uids       # list of strings, not a DataFrame
)

# Alternative: download_dicom_series has seriesInstanceUID as FIRST arg (different order!)
client.download_dicom_series(
    seriesInstanceUID=uids,      # FIRST arg here
    downloadDir="./data/lung_ct"
)

# Whole collection: downloadDir is still the FIRST positional argument
client.download_from_selection(downloadDir="./data/rider", collection_id="rider_pilot")

Both default to AWS; use source_bucket_location="gcs" for Google. In 0.12.5 the most-specific selector wins, so use SQL first for intersecting criteria and verify every requested UID exists.

Downloaded files are named <crdc_instance_uuid>.dcm, not by SOPInstanceUID. The DICOM UIDs are preserved inside the file metadata, not in the filename. Read DICOM headers for the series UID; crdc_instance_uuid is not a column of the series-level index.

idc download <collection|series-uid|manifest> --download-dir ./data does the same from a shell. See references/cli_guide.md for the dirTemplate hierarchy options (Python default: %collection_id/%PatientID/%StudyInstanceUID/%Modality_%SeriesInstanceUID; dirTemplate="" flattens), manifest downloads with resume, and dry-run size estimation.

3. Visualizing IDC images

viewer_url = client.get_viewer_URL(seriesInstanceUID=uid)        # one series
viewer_url = client.get_viewer_URL(studyInstanceUID=study_uid)   # all series in a study

Returns a browser URL — nothing is downloaded. The method selects OHIF v3 for radiology or SLIM for slide microscopy automatically. Viewing by study is useful when a single DICOM Study holds several Series (T1, T2, and DWI from one MRI session).

4. Licenses and citations — obligations, not optional steps

IDC data carries license terms and attribution requirements that follow it into any downstream publication or product, and neither is inferable from the pixel data. Check the license before use, and generate citations for whatever you download.

# License breakdown for a selection
licenses = client.sql_query("""
    SELECT DISTINCT collection_id, license_short_name,
           COUNT(DISTINCT SeriesInstanceUID) as series_count
    FROM index GROUP BY collection_id, license_short_name
""")

# Citations for the same selection you downloaded (APA by default)
for citation in client.citations_from_selection(collection_id="rider_pilot"):
    print(citation)

In the reviewed v24 snapshot, about 97% of IDC data by size is CC BY (commercial use allowed with attribution) and about 3% is CC BY-NC (non-commercial only). Licenses attach to series, not collections — 39 of 176 collections carry more than one — so check the selection you actually intend to use, and note that each component retains its license obligations.

Both tasks are available from all three access paths, so stay on whichever one the session is already using: idc-index as above, POST /v3/licenses and POST /v3/citations over REST, or the get_licenses and get_citations MCP tools. See references/licensing_and_citation.md for the full license inventory, all three routes, the citation formats (APA, BibTeX, CSL JSON, RDF Turtle), and what to include when publishing.

5. Reaching past the index

Pick the access path with the routing gate in Overview; Data Access Options above is the full routing table.

Before reaching for BigQuery (which needs a Google Cloud project and access), check whether a specialized index table already has the column you want: search client.indices_overview, then client.fetch_index(...) and query locally for free. Full instance metadata, per-segment rows, and SR measurement tables require BigQuery or its public Parquet exports; these are outside the compact idc-index tables.

Best Practices

Check schema before writing queries — Use client.get_index_schema('index') (reads cached metadata, no SQL executed) or client.indices_overview to see all available columns and their descriptions. The version-tracking columns series_init_idc_version and series_revised_idc_version in the main index table directly answer "what's new / when was this added" questions without touching prior_versions_index.
Use the index for IDC data content questions - Query the IDC index directly, via client.sql_query() locally or POST /v3/sql over HTTP. Web sources (release notes, blog posts, documentation pages) are frequently out of date and will produce incorrect answers. The index is the authoritative source; use it even when web search is available.
Verify the IDC data version at the start of a session - client.get_idc_version(), GET /v3/version, or the MCP get_idc_version tool, depending on the path in use (currently v24). For a stale local index, run scripts/check_version.py and use the upgrade command it prints
Check licenses and generate citations - Query license_short_name and respect CC BY vs CC BY-NC terms; use citations_from_selection() to produce citations from source_DOI for publications
Explore small, then commit - Use LIMIT (or a low max_rows) while exploring, and check collection size before downloading — some collections are terabytes. See references/cli_guide.md
Keep downloads reproducible - Organize with dirTemplate (e.g. %collection_id/%PatientID/%Modality) and save the Series UIDs or manifest behind any dataset you build

Troubleshooting

Issue: ModuleNotFoundError: No module named 'idc_index'

Cause: idc-index package not installed
Solution: If the task is read-only metadata, do not install it — use the REST API instead (Data Access Options). Otherwise run scripts/check_version.py and use the install command it prints, which targets the running interpreter and pins the vetted version. For data analysis also add pandas, numpy, and pydicom (tested with pandas>=1.5, numpy>=1.23, pydicom>=2.3)

Issue: Download fails with connection timeout

Cause: Network instability or large download size
Solution: Download in smaller batches (10-20 series); see references/cli_guide.md for --use-s5cmd-sync resume and retry guidance

Issue: BigQuery quota exceeded or billing errors

Cause: Project quotas, sandbox limits, or billing configuration prevent the query
Solution: Use idc-index mini-index for simple queries (no billing required), or see references/bigquery_guide.md for cost optimization tips

Issue: Series UID not found or no data returned

Cause: Typo in UID, data not in the current IDC version, or wrong field name
Solution: Test with LIMIT 5 first, check field names against client.indices_overview, and confirm the series is in the current version (some old data is deprecated)

Issue: Column not found in index table (e.g., SliceThickness, PixelSpacing, KVP, EchoTime, InjectedDose)

Cause: The index table contains series-level metadata only; modality-specific acquisition and reconstruction parameters live in dedicated tables (ct_index, mr_index, pt_index)

Solution: Search client.indices_overview for the column to find its table — the loop is under Finding which table contains a column in references/index_tables_guide.md — then fetch and join on SeriesInstanceUID:

client.fetch_index("ct_index")
result = client.sql_query("""
    SELECT i.SeriesInstanceUID, i.Modality, c.SliceThickness, c.KVP, c.PixelSpacing_row_mm
    FROM index i
    JOIN ct_index c USING (SeriesInstanceUID)
    WHERE i.collection_id = 'your_collection'
""")

Issue: Downloaded DICOM files won't open

Cause: Corrupted download, or an object type the viewer does not handle — SEG, RTSTRUCT, SR, and slide microscopy all need specialized tools
Solution: Inspect Modality, SOPClassUID, and transfer syntax with normal pydicom.dcmread; forced parsing is not validation. Check download integrity and decoder/viewer support before re-downloading; reserve force=True for diagnosed non-Part-10 inputs.

Resources

Reference guides and their decision triggers are listed in Quick Navigation above.

IDC Portal: https://portal.imaging.datacommons.cancer.gov/explore/
Documentation: https://learn.canceridc.dev/ — Tutorials: https://github.com/ImagingDataCommons/IDC-Tutorials
User Forum: https://discourse.canceridc.dev/ — idc-index: https://github.com/ImagingDataCommons/idc-index
indices_reference — external index-table docs (may be ahead of the installed version)
Citation: Fedorov, A., et al. "National Cancer Institute Imaging Data Commons: Toward Transparency, Reproducibility, and Scalability in Imaging Artificial Intelligence." RadioGraphics 43.12 (2023). https://doi.org/10.1148/rg.230180
Skill updates: releases page; watch the repository (Watch → Custom → Releases)

When it activates

Automatically, when a task matches. Use for accessing large-scale radiology (CT, MR, PET) and pathology datasets for AI training or research. No authentication required. Query by metadata, visualize in browser, check licenses.

About this skill

Automatic

Clinical / Healthcare pack

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v5 — 10/8/2026

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

k-dense-ai / scientific-agent-skills

skills/imaging-data-commons/SKILL.md

View on GitHub ↗

License This skill is provided under the MIT License. IDC data itself has individual licensing (mostly CC-BYImported 10/8/2026Version v3Bundled 14 files · scripts

Illustrative Want to see what k-dense-ai/imaging-data-commons 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

Imaging Data Commons

Overview

IDC MCP Server

When to Use This Skill

Quick Navigation

IDC Data Model

Index Tables

Available Tables

Joining Tables

Clinical Data Access

Data Access Options

Core Capabilities

1. Discovery — enumerate values before filtering on them

2. Downloading DICOM files

3. Visualizing IDC images

4. Licenses and citations — obligations, not optional steps

5. Reaching past the index

Best Practices

Troubleshooting

Resources

When it activates

SkillSafety

Source

Other skills for the same job

Related skills

More in the registry