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 job-diagnosegit 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/job-diagnose)<a href="https://agentmods.dev/skills/iowarp/clio-kit/job-diagnose"><img src="https://agentmods.dev/badge/skills/iowarp/clio-kit/job-diagnose/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/job-diagnose"><img src="https://agentmods.dev/badge/skills/iowarp/clio-kit/job-diagnose.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 | $0.00054 | $0.01235 |
| Opus 5.5 | $0.00022 | $0.00494 |
| Sonnet 5.5 | $0.00011 | $0.00247 |
| Haiku 4.5 | $0.00005 | $0.00123 |
Grade A, and why
job-diagnose 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 — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Job Diagnosis
A Darshan log records what a finished job did to the filesystem. This skill uses that evidence to develop and test a bottleneck hypothesis. It spans the profiler and the log tools, and the order is not optional.
Check the input, then load
Confirm that a genuine Darshan profile was supplied. An application text log
cannot substitute for it: do not send a .log file to the Darshan parser just
to discover that it is unsupported. With application logs only, use the log
tools and report which profiler evidence is missing.
clio-darshan:load_darshan_log parses the log and holds the result. Every other
darshan tool operates on what that call loaded — none of them take a log path.
Calling them first is an error, not an empty answer.
One log is loaded at a time. To look at a second run, load it; the first is
replaced. clio-darshan:compare_darshan_logs is the tool that holds two at once.
Steps
1. Load the log. clio-darshan:load_darshan_log.
2. Get the shape of the run. clio-darshan:get_job_summary — runtime, process
count, total I/O volume. Compute the crude number yourself: volume ÷ runtime. This is
average application throughput, not the fraction of time blocked on I/O.
Compare I/O time on the critical path and an equivalent filesystem baseline;
high throughput can coexist with an I/O-bound job.
3. Get the real metrics. clio-darshan:get_io_performance_metrics — bandwidth,
IOPS, request sizes. Request size is the field that usually explains everything;
see io-performance for what counts as bad.
4. Look at how the files were touched.
clio-darshan:analyze_file_access_patterns — read/write mix, sequential vs random.
5. Go down a layer, and pick the right layer.
clio-darshan:analyze_posix_operationsfor raw read/write syscall counts.clio-darshan:analyze_mpiio_operationsfor collective vs independent MPI-IO.
For an MPI code, inspect MPI-IO when that module is present. MPI programs can also issue POSIX I/O directly; missing MPI-IO counters do not establish that no I/O occurred. Avoid double-counting bytes observed at both layers.
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 · 106 lines · 54 tokens per session scan A 9286d4a71051
job-diagnose is a skill published in the GitHub repository iowarp/clio-kit (30 stars, last pushed 8d ago), licensed BSD-3-Clause. It adds 54 tokens to every session and 1,235 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
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…
compile-warning-handler
Compiler warning triage for HPC parallelization. Use when deciding whether to proceed with job execution after compile warnings.
bio-genome-assembly-assembly-polishing
Decides whether and how to polish a draft genome assembly to raise consensus accuracy (QV) with read-type-matched tools - Racon and medaka (ONT consensus), dorado polish, Polypolish and pypolca (Illumina, repeat-aware), Pilon (legacy short-read), NextPolish/NextPolish2, Hapo-G (haplotype-aware), ntEdit, and…
mxl-troubleshoot
This skill should be used when MaxwellLink simulations stall, diverge, or produce missing/incorrect observables; it provides a first-aid checklist for sockets, units, and driver configuration.
smiles-validation
Strict SMILES validation, structural comparison, and modification verification. Catches invalid LLM-generated molecules.
nonlinear-solvers
Select and configure nonlinear solvers for f(x)=0 or min F(x). Use for Newton methods, quasi-Newton (BFGS, L-BFGS), Broyden, Anderson acceleration, diagnosing convergence issues, choosing line search vs trust region, and analyzing Jacobian quality.