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 LearnPrompt/andrej-karpathy-skills --skill karpathy-supply-chain-hygienegit clone --depth 1 https://github.com/LearnPrompt/andrej-karpathy-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/learnprompt/andrej-karpathy-skills/karpathy-supply-chain-hygiene)<a href="https://agentmods.dev/skills/learnprompt/andrej-karpathy-skills/karpathy-supply-chain-hygiene"><img src="https://agentmods.dev/badge/skills/learnprompt/andrej-karpathy-skills/karpathy-supply-chain-hygiene/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/learnprompt/andrej-karpathy-skills/karpathy-supply-chain-hygiene"><img src="https://agentmods.dev/badge/skills/learnprompt/andrej-karpathy-skills/karpathy-supply-chain-hygiene.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 39 Skill grants unrestricted tool access without appropriate constraints. An agent with unfettered tool access can perform arbitrary actions including file modification, network requests, and code execution.Fix: Restrict tool access to only the tools required for the skill's stated purpose. Use an explicit allowlist rather than granting blanket access.
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 | $0.00111 | $0.01738 |
| Opus 5 | $0.00056 | $0.00869 |
| Sonnet 5 | $0.00022 | $0.00348 |
| Haiku 4.5 | $0.00011 | $0.00174 |
Grade A, and why
karpathy-supply-chain-hygiene 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 12d 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.
Unrestricted tool accesslowExcessive agency
A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.
3. Install-time code execution: does it run arbitrary code on install? Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 208 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill 11: Supply Chain & Security Hygiene（供应链安全卫生）
Source: https://x.com/karpathy/status/2036487306585268612 "Software horror: litellm PyPI supply chain attack" — ~28k likes
Core Principle
Every dependency is a door you didn't build, maintained by someone you don't know.
The litellm incident Karpathy flagged: a widely-used LLM library got compromised via its transitive dependencies. If your project pulled litellm, you were exposed — and you probably didn't even know litellm was in your stack until it was too late.
Karpathy's response: minimize deps, prefer simple implementations, audit everything.
The Full Dependency Audit
Run this for any project before deployment or after any dependency update:
Perform a full supply chain security audit for this project.
Dependency file contents:
[PASTE requirements.txt / package.json / Gemfile / go.mod]
For each direct dependency, analyze:
1. Maintainer health: single maintainer? Last commit? GitHub stars?
2. Transitive depth: how many packages does it pull in?
3. Install-time code execution: does it run arbitrary code on install?
4. Network calls on import: does importing trigger outbound connections?
5. Version pinning: are we using exact versions or ranges?
Risk assessment per package:
- HIGH RISK: (list criteria)
- MEDIUM RISK: (list criteria)
- LOW RISK: (list criteria)
Priority actions: [what to fix first]
The 5-Minute Pre-Install Check
Before adding any new package:
Security pre-check for: [PACKAGE NAME] v[VERSION]
1. PyPI/npm stats: weekly downloads? Known? Established?
2. GitHub: https://github.com/[owner/repo] — stars, last commit, open issues?
3. Transitive deps: run `pip show [package]` or `npm ls [package]` — how deep does it go?
4. setup.py / postinstall: does it run code on install?
5. Import-time behavior: what does `import [package]` actually do?
Risk verdict: SAFE / VERIFY_FURTHER / AVOID
Alternative: if risky, what can I use instead or implement myself?
Requirements File Hardening
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.
- 12d ago First seen · 208 lines · 111 tokens per session scan A 4c5b2b54ac16
karpathy-supply-chain-hygiene is a skill published in the GitHub repository LearnPrompt/andrej-karpathy-skills (97 stars, last pushed 2mo ago), licensed MIT. It adds 111 tokens to every session and 1,738 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…