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 mila-iqia/skills --skill mila-sarcgit clone --depth 1 https://github.com/mila-iqia/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/mila-iqia/skills/mila-sarc)<a href="https://agentmods.dev/skills/mila-iqia/skills/mila-sarc"><img src="https://agentmods.dev/badge/skills/mila-iqia/skills/mila-sarc/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/mila-iqia/skills/mila-sarc"><img src="https://agentmods.dev/badge/skills/mila-iqia/skills/mila-sarc.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00080 | $0.05735 |
| Opus 5 | $0.00040 | $0.02867 |
| Sonnet 5 | $0.00016 | $0.01147 |
| Haiku 4.5 | $0.00008 | $0.00573 |
Grade A, and why
mila-sarc scanned grade A with 1 finding 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 6d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
below is plain HTTP — curl, `requests`, or any client will do; there is nothing How it starts
The opening of the file, as written. The whole thing — 409 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fetching data from the SARC API
SARC serves a JSON API over HTTPS at https://sarc.mila.quebec. Everything
below is plain HTTP — curl, requests, or any client will do; there is nothing
to install.
Two families of endpoints matter:
| Prefix | What it gives you | Response |
|---|---|---|
/v0/… |
Raw records: slurm jobs, job series (jobs + computed usage/waste/RGU columns), users, clusters, GPU→RGU table, health checks | JSON |
/dash/metrics/… |
Pre-aggregated dashboard data: job counts, RGU usage, metric distributions, per-user/per-cluster rollups | JSON |
https://sarc.mila.quebec/dash/metrics itself is the human-facing dashboard
(HTML) — worth opening in a browser to see what the JSON endpoints feed.
The parameters below are enough to query the API directly — go ahead and call it.
If a call fails in a way this document doesn't explain, fetch
https://sarc.mila.quebec/openapi.json (public, no token; interactive at
/docs) — it's the authoritative description of what the server accepts.
Authentication
Every endpoint except /v0/health/list, /docs and /openapi.json requires a
bearer token:
Authorization: Bearer $SARC_TOKEN
Three environment variables carry everything you need. Check them first — an
unset SARC_TOKEN is the single most common reason requests come back
401 {"detail":"Authentication required"}:
test -n "$SARC_TOKEN" && echo "token: set" || echo "token: MISSING"
echo "user: ${SARC_USERNAME:-unset} admin: ${SARC_IS_ADMIN:-0}"
If SARC_TOKEN is empty, stop and give the user the snippet below to paste into
their ~/.bashrc (~/.zshrc works the same). Fill SARC_USERNAME in with your
best guess from the conversation if you have one (usually [email protected])
— otherwise leave the placeholder. Don't try to work around a missing token.
# --- SARC ---------------------------------------------------------------
# API token. Open https://sarc.mila.quebec/token in a browser, sign in with
# your Mila Google account, and copy the "refresh_token" value from the JSON
# page you land on. It is long-lived: treat it like a password, and never
# commit it.
export SARC_TOKEN='paste-the-refresh_token-value-here'
# Your @mila.quebec address — how SARC identifies you.
export SARC_USERNAME='[email protected]'
# 1 if your SARC account has the admin capability, 0 otherwise.
# Leave it at 0 if you don't know; you'd know if you had it.
export SARC_IS_ADMIN=0
# ------------------------------------------------------------------------
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.
- 6d ago First seen · 409 lines · 80 tokens per session scan A c2817ca37b91
mila-sarc is a skill published in the GitHub repository mila-iqia/skills (5 stars, last pushed 7d ago), licensed MIT. It adds 80 tokens to every session and 5,735 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-05.
Other skills, from other repositories
instrument-data-to-allotrope
Convert laboratory instrument output files (PDF, CSV, Excel, TXT) to Allotrope Simple Model (ASM) JSON format or flattened 2D CSV. Use this skill when scientists need to standardize instrument data for LIMS systems, data lakes, or downstream analysis. Supports auto-detection of instrument types. Outputs include full…
exploratory-data-analysis
Perform bounded, local exploratory analysis of explicitly supported scientific files. Use for redacted CSV/TSV/JSON profiles; optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection; missingness/leakage audits; outlier and transformation sensitivity; and rigorous EDA report scaffolds. Other domain…
matlab
Build, review, migrate, and safely plan MATLAB or GNU Octave numerical workflows, including arrays, tabular/time data, tests, projects, graphics, MAT files, and explicit Python interoperability.
phylogenetics
Build and analyze phylogenetic trees using MAFFT (multiple alignment), IQ-TREE 2 (maximum likelihood), and FastTree (fast NJ/ML). Visualize with ETE3 or FigTree. For evolutionary analysis, microbial genomics, viral phylodynamics, protein family analysis, and molecular clock studies.
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
mapping-to-snomed
Maps clinical concept spans extracted by OpenMed to SNOMED CT concepts through a USER-SUPPLIED terminology server (the user's own Ontoserver, Snowstorm, or UMLS/UTS), never a bundled vocabulary. Use when the user wants to code findings, disorders, procedures, body structures, or substances to SNOMED CT, run an ECL…