Observed since 20 Oct, 23:21 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
A framework for running AI agent workflows against code repositories.
Agents, Graphs, Loops
AGL runs Claude Code and Codex agents as workflows you write in Python. Each step runs an agent on a git branch AGL makes for the run, so your checkout stays as it is, and the run's work waits for you on one branch. AGL can use your subscription limits, so API keys are optional.
Install AGL with uv:
uv tool install agents-gl
Download the implement and review workflow. Claude implements your request, Codex reviews the change after your build runs, and Claude fixes what the review finds, for up to three rounds.
agl get jashioq/AGL-workflows/implement_and_review
Run it from your repository:
cd path/to/your/repo
agl run implement_and_review -n verbose-flag -r "Add a --verbose flag"
The first run registers the repository and stops to ask for its build command. Add it to the
project file the message names, ~/.agl/projects/<project>.toml, and run the same command again:
build = "npm run build && npm test"
build = "" skips the build. When the run finishes, its work is on the branch agl/verbose-flag.
AGL runs agents through Claude Code and Codex on your machine and uses their logins, so log in to the ones your workflows use before you run them. Implement and review uses both:
claude
codex login
claude walks you through its login. To install either tool, see
Getting started.
The full documentation is at agents-gl.org.
Found a bug, or have a request or a suggestion? Open an issue.
Download counts and popularity will appear here as the collector receives them.
| Version | Published | Direct dependencies | Open version |
|---|---|---|---|
| 0.1.3 | 2 Oct 2026 | 3 |
This describes our collected data; it is not a security assessment.
Observed since 20 Oct, 23:21 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
package:d604e279b506b8ae3c301b155173c1907cc9284b1ccbf893a475536cfca7553e