data-download¶
data-download fetches a selected Cell Painting dataset slice into a local folder.
It is the step that turns a download request into local files that later profiling or segmentation steps can read.
Purpose¶
Use this skill when you want:
local Cell Painting input files on disk
a small subset that can be used immediately in the next step
a clear record of which files were fetched
Main Outcome¶
This skill writes the requested Cell Painting files into a local download folder.
In the recorded demo run below, it downloaded the two metadata tables that the later demo steps use:
load_data.csvload_data_with_illum.csv
Inputs¶
This skill reads:
a project config such as
configs/project_config.demo.jsona data request, or a saved plan from data-plan-download
an optional output directory
In the demo run below, the request targeted dataset cpg0016-jump, source source_4, and a small metadata subset for BR00117035.
Outputs¶
downloads/load_data.csvThe field-level metadata table used by downstream demo steps.downloads/load_data_with_illum.csvThe same metadata table with illumination-related path columns.download_plan.jsonThe saved request that was executed for this download.downloads/download_manifest.jsonThe file-level record of what was downloaded.pipeline_skill_manifest.jsonThe run record for this skill invocation.
Agent Demo¶
This page is based on a real local OpenClaw run against the repository demo config:
session id:
datadownload-local-v1turn timestamp:
2026-05-05 13:40 UTCmodel:
vibe/gpt-5-mini
Request¶
I want a tiny local demo subset of the Cell Painting inputs so I can test the next step. Please download the BR00117035 metadata subset from the demo dataset and then tell me what became available locally.
Structured Trace¶
user_input:
I want a tiny local demo subset of the Cell Painting inputs so I can test the next step. Please download the BR00117035 metadata subset from the demo dataset and then tell me what became available locally.
planner_step_1:
intent:
Load the documented Cell Painting skill routing before choosing a data-access execution skill.
tool_calls:
- read("integrations/openclaw/autodl/workspace/skills/cellpaint-pipeline/SKILL.md")
observed_result:
- the workspace skill file loaded successfully
- the routing guidance identified `data-download` as the public skill for fetching local input files
planner_step_2:
intent:
Confirm the current skill catalog and inspect the documented description for the selected skill.
tool_calls:
- exec("cd /root/pipeline/CellPainting-Claw && /root/autodl-tmp/miniconda3_envs/lyx_env/bin/cellpainting-skills list && echo '---DESCRIBE---' && /root/autodl-tmp/miniconda3_envs/lyx_env/bin/cellpainting-skills describe --skill data-download")
observed_result:
- the public skill catalog listed `data-download`
- the skill description confirmed that this step downloads one configured dataset slice into a local cache directory
planner_step_3:
intent:
Locate the documented demo subset and run the actual metadata download into the named demo output directory.
tool_calls:
- exec("cd /root/pipeline/CellPainting-Claw && /root/autodl-tmp/miniconda3_envs/lyx_env/bin/cellpainting-skills run --config /root/pipeline/CellPainting-Claw/configs/project_config.demo.json --skill data-download --dataset-id cpg0016-jump --source-id source_4 --subprefix workspace/load_data_csv/2021_04_26_Batch1/BR00117035 --output-dir /root/pipeline/CellPainting-Claw/demo/workspace/outputs/agent_demo_data_download")
- process.poll("nova-crest")
observed_result:
- skill_key = data-download
- resolved prefix = cpg0016-jump/source_4/workspace/load_data_csv/2021_04_26_Batch1/BR00117035/
- matched files = 2
- downloaded files = 2
- the download completed successfully
planner_step_4:
intent:
Inspect the written output tree and report exactly which files were fetched.
tool_calls:
- exec("cd /root/pipeline/CellPainting-Claw && find demo/workspace/outputs/agent_demo_data_download -maxdepth 3 -type f | sort")
observed_result:
- download_plan.json
- downloads/download_manifest.json
- downloads/load_data.csv
- downloads/load_data_with_illum.csv
- pipeline_skill_manifest.json
final_answer_snapshot:
The demo metadata subset downloaded successfully and made two local metadata tables available for the next step.
Observed Result¶
The recorded run downloaded one small Cell Painting metadata subset from:
dataset
cpg0016-jumpsource
source_4gallery prefix
cpg0016-jump/source_4/workspace/load_data_csv/2021_04_26_Batch1/BR00117035/
The saved run record showed:
matched files:
2downloaded files:
2
The fetched files were:
load_data.csvload_data_with_illum.csv
Recorded Reply¶
I downloaded the requested demo metadata subset successfully.
Source used:
- dataset: `cpg0016-jump`
- source: `source_4`
- gallery prefix: `cpg0016-jump/source_4/workspace/load_data_csv/2021_04_26_Batch1/BR00117035/`
What is now available locally:
- `load_data.csv`
- `load_data_with_illum.csv`
Summary:
- matched files: `2`
- downloaded files: `2`
This was a metadata-only download. In practical terms, you now have the field-level image metadata needed to start the next profiling or segmentation step.