read-unfamiliar-code

Fetches upstream package source and maps it into a queryable call graph for code comprehension.

Updated Jun 9, 2026
One-click install
npx skills add https://github.com/timikalo7/Execute --skill read-unfamiliar-code-timikalo7
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: read-unfamiliar-code
Source: https://github.com/timikalo7/Execute/tree/main/.claude/skills/read-unfamiliar-code
Command: npx skills add https://github.com/timikalo7/Execute --skill read-unfamiliar-code-timikalo7

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires opensrc, graphifyy.

What problem does it solve? When working with a dependency, inherited project, or open-source repo you did not write, READMEs and docs describe intent rather than actual behavior, leading to wrong-API usage and hallucinated signatures. This Skill pulls the real upstream source and maps it into a call graph so you read the files that actually matter. ## Core Features & Use Cases - Real Source Fetching: Uses opensrc to download the actual GitHub source tree behind npm, PyPI, or crates.io packages at the exact version you depend on. - Call Graph Mapping: Uses graphify to build a local AST-based call/containment graph with zero tokens, no LLM, and no API key, producing a queryable graph.json and a human-readable report. - Targeted Reading Workflow: Query symbols, call paths, and relationships with graphify explain/path/query before opening files, so you read twenty relevant files instead of two thousand. - Use Case: Before modifying an inherited codebase or depending on a library whose behavior matters, run opensrc path zod then graphify update on the result, query the graph for the symbols you care about, and read only those files. ## Quick Start Ask the agent to pull the real source of the library you are about to use with opensrc, map it with graphify, and explain how a specific function is called before you change any code.

Frequently Asked Questions about read-unfamiliar-code

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I understand a codebase I did not write?

Fetch the real upstream source with opensrc, then build a call graph with graphify to see which files matter. Query symbols with graphify explain or path, then read only the files the graph points at instead of the whole tree.

How to get the actual source code behind an npm or PyPI package?

Run opensrc path followed by the package name, such as opensrc path zod or opensrc path pypi:requests. It clones the actual GitHub repository at the version you depend on into a local cache and prints the path.

Does graphify need an API key or LLM to analyze code?

No, code extraction in graphify is fully local tree-sitter AST parsing with no LLM, no API key, and no network. Only semantic extraction from docs, PDFs, images, or video requires an LLM backend key.

Why does opensrc fail with a request error in a proxied environment?

The opensrc binary cannot reach registries on the direct-egress path set by no_proxy. Clear it with no_proxy="" NO_PROXY="" before the command so traffic goes through the agent proxy.

What are the limitations of graphify call graphs?

Multiple call sites from one function collapse into a single edge, so the graph is an index, not a census. INFERRED edges are low-confidence guesses, and graphify explain exits 0 even when a node does not exist.

When should I not use source-graph analysis for a library?

Skip it for small files you can just read directly, and for libraries where the docs are the authority, such as a public HTTP API spec. The graph is for trees too big to hold in context.