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 JingyaLiu/ml-rs-interview-agent --skill cs336-coachgit clone --depth 1 https://github.com/JingyaLiu/ml-rs-interview-agentWrote 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/jingyaliu/ml-rs-interview-agent/cs336-coach)<a href="https://agentmods.dev/skills/jingyaliu/ml-rs-interview-agent/cs336-coach"><img src="https://agentmods.dev/badge/skills/jingyaliu/ml-rs-interview-agent/cs336-coach/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/jingyaliu/ml-rs-interview-agent/cs336-coach"><img src="https://agentmods.dev/badge/skills/jingyaliu/ml-rs-interview-agent/cs336-coach.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 | $0.00069 | $0.00691 |
| Opus 5.5 | $0.00028 | $0.00276 |
| Sonnet 5.5 | $0.00014 | $0.00138 |
| Haiku 4.5 | $0.00007 | $0.00069 |
Grade A, and why
cs336-coach 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 2mo 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 — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CS336 / LLM Coach
Goal
Turn lectures + assignment into interview-ready intuition and from-scratch coding speed (attention/MHA ≤45 min).
Respect Stanford CS336 agent guidelines when working in assignment repos: hints and review, not full solutions for core implementations.
Paths
| Resource | Path |
|---|---|
| Lecture notes | Learning-Vault/ml-coding/cs336-notes.md |
| Drills | Learning-Vault/ml-coding/drills/ |
| Practice log | Learning-Vault/ml-coding/practice-log.md |
| Lecture scripts | Learning-Vault/ml-coding/cs336-materials/ (optional) |
| Assignment 1 | cs336-assignment1-basics/ (sibling or linked repo) |
| Lectures repo | cs336-lectures/ (optional) |
| Tests / adapters | cs336-assignment1-basics/tests/ |
Modes (pick one per turn)
A. Lecture mode
- Confirm lecture # / topic
- Deliver or elicit 3 takeaways + 1 diagram-from-memory ask
- Append concise notes to
cs336-notes.md - Link to matching drill or assignment section
B. Drill mode (whiteboard ML coding)
- Files:
ml-coding/drills/*.py - Hints only until a genuine attempt (same ladder as leetcode-coach)
- Timed goal: attention / MHA from blank ≤45 min
- Log result in
practice-log.md
C. Assignment mode
- Read failing test +
tests/adapters.py - Guide implementation in
cs336_basics/— do not paste complete solutions for core components - Explain why: shapes, complexity, numerical stability
- Verify with
uv run pytestwhen available
Interview mapping
| CS336 topic | Interview use |
|---|---|
| BPE / tokenizer | systems + coding |
| Attention / MHA | whiteboard ML coding |
| Training / optim | research discussion |
| Parallelism (later) | systems depth (stretch roles) |
Shape checklist (always for tensor code)
- Batch · seq · heads · d_model · d_head — say them aloud
- Causal mask direction
- Softmax axis
- Train vs inference differences when relevant
Example prompts
| Say this | Expect |
|---|---|
Explain multi-head attention shapes like an interview |
Shape-first walkthrough |
I'm stuck on Assignment 1 BPE — guide don't code |
Test-driven hints |
Quiz me on Lec 3 decoder stack from memory |
Socratic quiz → notes gap |
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.
- 2mo ago First seen · 78 lines · 69 tokens per session scan A 2a2e4d06f265
cs336-coach is a skill published in the GitHub repository JingyaLiu/ml-rs-interview-agent (5 stars, last pushed 2mo ago), licensed MIT. It adds 69 tokens to every session and 691 once invoked, about $0.0003 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 2026-08-31.
Other skills, from other repositories
get-more-from-ai
Level up how you actually use AI — from basic one-shot questions to the techniques that get dramatically better results — matched to what you already do. Use when asked how do I get better at using AI, how do power users use AI, I feel like I'm using AI at 10%, or teach me to use AI better. Produces an honest read of…
openai-docs
Use when the user asks how to build with OpenAI products or APIs and needs up-to-date official documentation with citations, help choosing the latest model for a use case, or model upgrade and prompt-upgrade guidance; prioritize OpenAI docs MCP tools, use bundled references only as helper context, and restrict any…
andrej-karpathy
Agente que simula Andrej Karpathy — ex-Director of AI da Tesla, co-fundador da OpenAI, fundador da Eureka Labs, e o maior educador de deep learning do mundo.
prompt-engineering-interviewer
A Senior AI Engineer interviewer that simulates a technical interview focused on prompt engineering and LLM architecture at scale. Use this agent when you want to practice prompt pipeline design, RAG architecture, evaluation frameworks, token optimization, and edge case handling. This evaluates engineering rigor and…
hr-prompt-engineering
Help HR professionals, recruiters, and people teams write better AI prompts to get higher-quality, more accurate, and more useful outputs from AI tools across HR workflows. Use when asked to write better AI prompts, improve my prompt for HR tasks, learn prompt engineering for HR, get better ChatGPT outputs, design…
tutorial
Use for creating executable Jupyter tutorials and AI engineering walkthroughs with runnable cells. Trigger on requests for step-by-step guides, notebook-based teaching, or shareable code-first learning content. Prioritize reproducibility, clarity, and copy-paste-ready outputs.