Running Pipelines

repos run executes a script inside each repository listed in your repos.list file. This lets you run a build, test, or any other task across your entire workspace with a single command.

Basic Usage

By default, repos run looks for a file called run.sh in each cloned repository and executes it:

repos run

Options

Flag	Description
--script <file>	Run a different script (default: run.sh)
-f <file>	Use a different list file (default: repos.list)
--include <repos>	Comma-separated list of repositories to include
--exclude <repos>	Comma-separated list of repositories to skip
--continue-on-error	Keep running even if a repository fails
--dry-run	Print what would be run without executing anything
--ensure-setup	Clone repositories before running the pipeline

# Run a custom script in each repo
repos run --script build.sh

# Only run in specific repos
repos run --include "backend,frontend"

# Skip a repo
repos run --exclude "docs"

# Continue even if one repo fails, then report all results
repos run --continue-on-error

# Preview without running
repos run --dry-run

# Make sure repos are cloned before running
repos run --ensure-setup

Writing a Pipeline Script

Each repository should contain the script that repos run will call. The script receives no arguments by default, but can read environment variables or accept flags if needed. A minimal run.sh:

#!/usr/bin/env bash
set -e

# Install dependencies
Rscript -e "renv::restore()"

# Run analysis
Rscript analysis.R

R and Python APIs

For the R or Python API, see the language-specific wrapper pages:

→ R package
→ Python package