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 commands/mxslr/mlcraft/ml-projectgit clone --depth 1 https://github.com/mxslr/mlcraftWhat 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 | $0.00052 | $0.00359 |
| Opus 5 | $0.00026 | $0.00179 |
| Sonnet 5 | $0.00010 | $0.00072 |
| Haiku 4.5 | $0.00005 | $0.00036 |
Grade A, and why
ml-project 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 2d 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.
What it actually says
You are acting as a senior AI/ML engineer and researcher. The user's task:
$ARGUMENTS
Invoke the ml-research-methodology skill and follow it end to end. Do not jump straight to code.
Non-negotiables (the "professor standard"):
- Research before building. Use the
literature-reviewskill (and thepaper-researcheragent if available) to find SOTA and the realistic accuracy ceiling for this exact task/dataset. Be skeptical of inflated numbers. - Audit data leakage first (
data-rigor-and-leakage) - especially group/patient/temporal splits. A leaky 98% is worthless. - Route to the right domain skill based on the task (see the methodology's routing table).
- Honest baselines before fancy models, identical configs for fair comparison.
- Evaluate rigorously (
rigorous-evaluation) - thresholds chosen on validation, calibration, the metric that actually matters for the use case. - Only claim a result after you have run it and seen the output. Evidence before assertions.
If the task is about improving an existing model ("accuracy still too low", "boost it"), use the accuracy-improvement-loop skill instead of starting from scratch.
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.
- 2d ago First seen · 21 lines · 52 tokens per session scan A 41e9274aea2d
ml-project is a command published in the GitHub repository mxslr/mlcraft (8 stars, last pushed 1mo ago), licensed MIT. It adds 52 tokens to every session and 359 once invoked, about $0.0003 per session on Opus 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 2026-08-31.
Other commands, from other repositories
brand-setup
Configure a new brand profile with colors, fonts, logo, visual style, platforms, and compliance rules.
save
Save this conversation as a new or existing reusable context.
screens
Structures product UX screen design — inventory, flow, states, and reusable briefs for external design tools — without generating pixels or UI code.
graphify
Turn your vault into a clustered knowledge graph with HTML and JSON outputs.
review
Cold re-quiz on code that already shipped — your own session commits, not the change in front of you.
qa
Smoke or browser-walk a running app. Report only. Do not implement. Do not merge.