agent-eval
colbymchenry1.7k installs
See whether CodeGraph saves effort on real code.
What it does
Benchmarks CodeGraph against plain code search on a selected repository and reports comparative effort, cost, and answer correctness.
When to use it
Use when asked to test, benchmark, audit, or validate a local or published CodeGraph version against a language's repository.
How to use it
Give it a benchmark request from the codegraph repository; it asks you to choose a version, language, repository, and harness, then reports the comparison.
What you provide
- An existing project
Uses
Good to know
- starts helper agents, number unstated
- cannot be undone: .codegraph
Latest review
What you need · 4
Requires the claude CLI, already logged in, to run the benchmark agents.
Requires an account with which the claude CLI is already logged in.
Access · 3
This skill
.codegraph
Write
Deletes the existing CodeGraph index (irreversible)
GitHub
Read
Downloads selected corpus repositories