FORGE

Agents don't just run. They evolve.

API Offline
Back to Dashboard

Create Evolution Experiment

Define agent goal, capabilities, and target benchmark

Repository Inspector
List, explore, and read repo files sandboxed to workspace
repository
File Editor
Create, patch, and replace code files with diff tracing
file_editor
Restricted Shell
Execute build and analysis commands inside workspace
shell
Automated Test Runner
Execute pytest suites and extract structured failure diagnostics
test_runner
Code Search
Grep and regex search codebase for symbol references
search
Cancel