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 skills add iowarp/clio-kit --skill large-data-readgit clone --depth 1 https://github.com/iowarp/clio-kitWrote 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/iowarp/clio-kit/large-data-read)<a href="https://agentmods.dev/skills/iowarp/clio-kit/large-data-read"><img src="https://agentmods.dev/badge/skills/iowarp/clio-kit/large-data-read/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/iowarp/clio-kit/large-data-read"><img src="https://agentmods.dev/badge/skills/iowarp/clio-kit/large-data-read.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00048 | $0.01516 |
| Opus 5.5 | $0.00019 | $0.00606 |
| Sonnet 5.5 | $0.00010 | $0.00303 |
| Haiku 4.5 | $0.00005 | $0.00152 |
Grade A, and why
large-data-read 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 9d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dataset Reading
Avoid transferring a whole dataset into the agent context to compute one number. Prefer a server-side reduction when it meets the requested coverage and accuracy. An exact reduction may still need to read every value.
Work out the size first
For a supplied HDF5 file, resolve its path and call open_file directly. Resource
discovery scans configured directories; an empty listing does not mean that the supplied file is missing.
From get_shape and get_dtype: multiply the dimensions, multiply by the item
size. A (2000, 2000, 500) array of float64 is 16 GB. Use it to choose a bounded read, and allow for temporary arrays and conversion
overhead. Shape and dtype require separate tool calls.
If you want a statistic, do not read the data
| Question | Tool |
|---|---|
| Mean/min/max over all elements of each HDF5 dataset | clio-hdf5:hdf5_aggregate_stats |
| Aggregate over a Parquet column | clio-parquet:aggregate_column_tool |
| Discover structure across many HDF5 files | clio-hdf5:hdf5_parallel_scan |
hdf5_aggregate_stats reduces all elements of each selected dataset; it has no
column or axis selector. Do not use its scalar mean as one column's mean in a
multicolumn array, especially when columns have different units. Select a tool
that implements the requested column reduction, or report the missing capability.
Open the HDF5 file before requesting aggregate statistics. For datasets larger
than 500 MiB, hdf5_aggregate_stats may use a strided sample. Read the response's
SAMPLED or FULL DATA label, processed/total element counts and coverage.
Sample sum/count/min/max describe selected values only, not whole-dataset
totals or bounds. Cross-dataset aggregation is omitted when any result is sampled.
Striding can miss patterns; it is not a random representative sample. Older
server versions may omit these labels, so verify coverage against the full shape.
One- and multidimensional arrays use different stride logic. Use the actual
response's counts and coverage; before a call, leave those quantities unverified.
Multidimensional sampling also does not guarantee a small allocation.
For exact results, use a separately verified chunked calculation with complete
coverage or report that the available tool cannot establish the exact answer.
Do not request thousands of hdf5_stream_data summaries to reconstruct an exact
mean: they are rounded summaries, not lossless sums. Establish that a suitable
local reader is available before choosing a fallback. Inspect its actual callable
and selection/coverage behavior; an importable library or a source file containing
the existing aggregate does not establish that a chunked column reader exists.
What ships with it
1 file 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.
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.
- 9d ago First seen · 121 lines · 48 tokens per session scan A 6040be85dc19
large-data-read is a skill published in the GitHub repository iowarp/clio-kit (30 stars, last pushed 8d ago), licensed BSD-3-Clause. It adds 48 tokens to every session and 1,516 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-08.
Other skills, from other repositories
bio-ortholog-inference
Pull pre-computed ortholog calls from public databases (OrthoDB, Ensembl Compara, OMA browser, eggNOG, PANTHER, KEGG Orthology, HomoloGene) via their REST APIs. Use when orthologs are already curated upstream, when the question is "what is the X ortholog of Y" rather than "how to infer orthology de novo", when…
admet_genetic
ADMET-guided genetic molecule optimization workflow from seed SMILES; use when the agent needs to build or run an RDKit/SA-Score/ADMET-AI GA pipeline for molecule optimization, enforce molecule lineage logs, render optimization-history HTML dashboards, and write candidate triage reports.
paper-figures
Use this skill to produce standalone, publication-ready PNG graphics and reproducible matplotlib scripts from tabular data (CSVs or DataFrames). It renders numerical data into formal scientific visualizations—including scatter, line, bar, pie, ring, bubble, tornado, KDE, violin, box, heatmap, histogram, and area…
paper-planning
Guides pre-writing planning for academic papers with 4 structured steps: story design (task-challenge-insight-contribution-advantage), experiment planning (comparisons + ablations), figure design (pipeline + teaser), and 4-week timeline management. Includes counterintuitive planning tactics (write a mock rejection…
paper-review
Guides self-review of YOUR OWN academic paper before submission with adversarial stress-testing. Core method: three passes (adversarial deep read; 5-aspect checklist — contribution sufficiency, writing clarity, results quality, testing completeness, method design; mechanical consistency scans and experimental protocol…
experiment-craft
Use this skill when the user wants to debug, diagnose, or systematically iterate on an experiment that already exists, or when they need a structured experiment log for tracking runs, hypotheses, failures, results, and next steps during active research. Apply it to underperforming methods, training that will not…