karpathy-graph skillA
karpathy-graph is agent-read markdown (skill) from andersonlimahw/lemon-ai-hub: Build typed, provenance-tracked knowledge graphs from any context — documents, codebases, incidents, research — so one or many AI agents share durable memory instead of copying transcripts. Use when the user says "build a knowledge graph", "criar grafo", "map this codebase/domain", "graph these documents", "extract entities and relations", "ground this claim", "visualize the graph", or asks how facts, sources and experiments connect. Every claim carries a source, every merge is reversible, every.
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
# Karpathy Graph You are building the **graph plane** of an agentic system: the layer that remembers what the context window forgets. > The agent forgets, the graph does not. This skill turns any context — a folder of documents, a codebase, an incident, a research thread — into a typed graph that both machines and humans can read. One agent can build it. Many agents can build it concurrently. A human opens `graph.html` and sees the same thing the agents queried. ## The one rule that matters **Never write an unsourced fact.** Every `Claim` node either points at a `Source`, or is explicitly marked `inference=true`. The engine enforces this — `kgraph validate` fails otherwise. If you cannot cite where a fact came from, it does not enter the graph. ## Before you build: does this need a graph? A graph earns its cost when connected queries, evolving relations, provenance, or shared world state are central. Skip it when tasks are independent, answers come from one document, no cross-session state is needed, or a flat table would answer every question. Say so plainly and stop — do not build a graph just because agents are involved. …
Read the whole file at its exact version.
How to install
mdr add andersonlimahw/lemon-ai-hub/karpathy-graph@v1.0.0mdr add andersonlimahw/lemon-ai-hub/karpathy-graph@sha256:dd105c0ff5bdec62Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_wmzji45csavvdbfi)
1 badge views in 30 days
Versions
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (9160 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
andersonlimahw/lemon-ai-hub · 14 stars · license none · pushed 2026-09-20 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_wmzji45csavvdbfi GET https://markdownregistry.com/api/v1/resolve?ref=andersonlimahw/lemon-ai-hub/karpathy-graph GET https://markdownregistry.com/api/v1/blob/dd105c0ff5bdec624c4aba45aab91753977bad18c0e2ea1b1c8525ced23a57ea
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.