Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/practicalswan/agent-skills/accelerated-computing-cudfnpx skills add PracticalSwan/agent-skills --skill accelerated-computing-cudfgit clone --depth 1 https://github.com/PracticalSwan/agent-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/practicalswan/agent-skills/accelerated-computing-cudf)<a href="https://agentmods.dev/skills/practicalswan/agent-skills/accelerated-computing-cudf"><img src="https://agentmods.dev/badge/skills/practicalswan/agent-skills/accelerated-computing-cudf.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00056 | $0.02895 |
| Opus 5 | $0.00028 | $0.01448 |
| Sonnet 5 | $0.00011 | $0.00579 |
| Haiku 4.5 | $0.00006 | $0.00290 |
Grade A, and why
accelerated-computing-cudf scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
88% identical to accelerated-computing-cudf — 65 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 247 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cuDF & dask-cuDF Implementer's Guide
Compatibility
- Release tracked by this skill: 26.04.
- Requires NVIDIA Volta or newer on CUDA 12, or Turing or newer on CUDA 13. Release 26.04 supports CUDA 12.2-12.9 with driver 535+ or CUDA 13.0-13.1 with driver 580+, and Python 3.11-3.14. cuDF sweet spot: >100K rows.
Naming
Use NVIDIA library-first wording in user-facing answers. Keep literal RAPIDS/rapidsai URLs, package names, and release metadata when citing sources.
Role
You are a cuDF expert helping an implementer work with GPU DataFrames. The user understands pandas and their data — your job is to get them to correct, fast GPU code with minimal friction. Choose the path from the user's intent: cudf.pandas for broad compatibility or minimal-change acceleration, explicit cuDF for named DataFrame migrations, hot ETL paths, and parity-sensitive work. Treat source schema, row counts, null placement, ordering, and numeric tolerances as user-visible behavior.
Critical Rules
- Choose the right cuDF path. Use
cudf.pandasfor broad compatibility or minimal-change acceleration. Use explicit cuDF when the user asks to migrate DataFrame code, inspect parity, optimize a visible ETL hot path, or control unsupported operations. - Size gate: 100K rows minimum. Below that, GPU transfer overhead usually beats the speedup; use small data for correctness and benchmark larger working sets for performance.
- Keep conversions at boundaries. Use
.to_pandas(),.values, or.numpy()for display, plotting, CPU-only libraries, or final output boundaries. Keep intermediate ETL data on GPU. - Float32 is your friend. cuDF operations on float64 are slower; cast early when precision allows.
- Validate semantics on representative slices. For null handling, joins, time series, reshape, or grouped logic, keep a small pandas reference path and compare shape, labels, null counts, ordering, and representative values before claiming parity.
- For data > GPU memory, move to dask-cuDF with
enable_cudf_spill=True. Seereferences/dask-cudf-patterns.md.
What ships with it
34 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- BENCHMARK.md 4.2 KB
- CHANGELOG.md 4.0 KB
- evals/evals.json 16 KB
- evals/files/cudf-apply-udf/code/generate_data.py 1.5 KB runs code
- evals/files/cudf-apply-udf/code/udf_pipeline.py 5.3 KB runs code
- evals/files/cudf-csv-etl/code/etl_pipeline.py 2.6 KB runs code
- evals/files/cudf-csv-etl/code/generate_data.py 1.2 KB runs code
- evals/files/cudf-groupby-agg/code/generate_data.py 1.4 KB runs code
- evals/files/cudf-groupby-agg/code/groupby_analysis.py 4.3 KB runs code
- evals/files/cudf-multi-join/code/generate_data.py 2.0 KB runs code
- evals/files/cudf-multi-join/code/multi_join.py 3.6 KB runs code
- evals/files/cudf-native-stream-handoff-boundary/code/run_smoke.sh 464 B runs code
- evals/files/cudf-native-stream-handoff-boundary/code/threaded_handoff.cu 3.9 KB
- evals/files/cudf-native-stream-handoff-boundary/NOTICE.md 344 B
- evals/files/cudf-null-handling/code/generate_data.py 2.1 KB runs code
- evals/files/cudf-null-handling/code/null_pipeline.py 4.7 KB runs code
- evals/files/cudf-parquet-io/code/generate_data.py 2.0 KB runs code
- evals/files/cudf-parquet-io/code/parquet_pipeline.py 4.3 KB runs code
- evals/files/cudf-pivot-melt/code/generate_data.py 1.4 KB runs code
- evals/files/cudf-pivot-melt/code/reshape_analysis.py 4.6 KB runs code
- evals/files/cudf-string-ops/code/clean_contacts.py 3.4 KB runs code
- evals/files/cudf-string-ops/code/generate_data.py 2.7 KB runs code
- evals/files/cudf-timeseries-resample/code/generate_data.py 1.5 KB runs code
- evals/files/cudf-timeseries-resample/code/timeseries_analysis.py 4.1 KB runs code
- evals/files/cudf-window-functions/code/generate_data.py 1.4 KB runs code
- evals/files/cudf-window-functions/code/window_analysis.py 5.1 KB runs code
- evals/files/negative-deep-learning-training/code/train.py 1.4 KB runs code
- evals/files/source-cudf-null-fillna-semantics/code/null_cleanup.py 1.4 KB runs code
- evals/files/source-cudf-null-fillna-semantics/NOTICE.md 342 B
- references/api-patterns.md 6.9 KB
- references/cudf-pandas-accelerator.md 3.6 KB
- references/dask-cudf-patterns.md 7.0 KB
- skill-card.md 4.0 KB
- skill.oms.sig 12 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday Changed 3741f743669b
- 5d ago First seen · 247 lines · 56 tokens per session scan A 63325e06664f
accelerated-computing-cudf is a skill published in the GitHub repository PracticalSwan/agent-skills (13 stars, last pushed 4d ago), licensed MIT. It adds 56 tokens to every session and 2,895 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. It is 88% identical to accelerated-computing-cudf, differing in 65 lines, and is treated as a copy.
Other skills, from other repositories
neo4j-knowledge-graph
Use when designing, importing, querying, or modernizing Neo4j knowledge graphs from CSV, Excel, pandas, Cypher, py2neo, the official neo4j Python driver, vector indexes, or GraphRAG workflows.
geopandas
Use when performing vector spatial data analysis in Python — reading/writing shapefiles, spatial joins, overlay operations, choropleth maps. GeoPandas: extends pandas DataFrames with geometry columns for Pythonic spatial analysis.
python-panel-data
Panel data analysis with Python using linearmodels and pandas.
rapids-first
Use only when the user explicitly mentions "rapids-first" or "RAPIDS first", to prefer the RAPIDS GPU stack (cudf, cuml, cupy, cupyx) and its zero-code accelerators over pandas, scikit-learn, numpy and scipy when writing Python data-science, ML, ETL or numerical code.
python-panel-data
Panel data analysis with Python using linearmodels and pandas.
data-analysis
Structured data analysis workflow from raw data to shareable insights.