keep_github_workflows_active

CI cadence: The GitHub Actions CI workflow runs automatically every Monday at 07:00 UTC, in addition to push and pull request triggers.

Overview

keep_github_workflows_active is an automation helper that tends to every GitHub Actions workflow in a user's repositories. It exists to remove the routine maintenance burden that accrues when repositories go quiet and Actions start to disable schedules automatically.

The automation focuses on two responsibilities:

1. Keep all workflows active – iterate every repository owned by the configured user and flip disabled workflows back to the active state so cron-based automation continues to run.
2. Delete stale workflow runs – retain an operator-defined number of the freshest runs per workflow and delete older artifacts so storage quotas and repository hygiene remain under control.

Both behaviors run as part of repository automation (for example on a nightly schedule) once the required GitHub token is in place.

Prerequisites

Generate an access token

The automation requires a fine-grained personal access token with repository scope. You can create one by visiting the GitHub self-service page:

• Generate fine-grained personal access token

When configuring the token, grant at least the following permissions:

• Actions: Read/Write (needed to toggle workflow states and delete runs).
• Metadata: Read (lets the script discover repositories you own).

Tokens expire automatically. Renew them before the expiry hits to keep the automation alive. The previous test token that powered the repository expired on January 19, 2025, so any current setup must already use a newer credential.

Store the token as a repository secret

Add the token to the repository's secrets so GitHub Actions can inject it at runtime:

• Repository secrets

Define the following secrets:

• SECRET_GITHUB_OWNER – GitHub username whose repositories should be scanned.
• SECRET_GITHUB_TOKEN – the fine-grained personal access token you generated.

Keeping the credentials inside secrets ensures the nightly cleanup workflow can authenticate without hard-coding sensitive values in the repository.

For local development you can place the same keys inside a .env file at the repository root. The automation reads from the environment first and falls back to .env when running outside GitHub Actions.

Note: the included tests and CLI commands exercise the live GitHub API. When SECRET_GITHUB_OWNER and SECRET_GITHUB_TOKEN point at a real account, running make test (or the workflow maintenance commands directly) will re-enable workflows and prune old workflow runs across the configured repositories.

Installation

Install the package locally when you want to run the CLI helpers or contribute changes:

pip install keep_github_workflows_active

Alternative flows (pipx, uv, editable installs, or source builds) are described in INSTALL.md. All supported options expose both the keep_github_workflows_active and keep-github-workflows-active entry points on your PATH.

Python 3.13+ baseline

• The package targets Python 3.13 and newer. Older interpreters are no longer supported; the codebase relies on modern conveniences such as Path.unlink(missing_ok=True).
• Runtime modules now use native Python 3.13 type syntax (e.g., list[str], Sequence[str] | None) and drop legacy compatibility helpers.
• Runtime dependencies include rich-click for the CLI surface and lib_cli_exit_tools for consistent exit handling. Development extras pin the tooling stack (pytest, ruff, pyright, bandit, build, twine, codecov-cli, pip-audit, textual, import-linter) to their latest major releases.
• Development dependencies were refreshed to the latest stable releases (ruff>=0.14.1, textual>=6.4.0).
• Continuous integration runs across GitHub's hosted runners (ubuntu-latest, macos-latest, windows-latest) on CPython 3.13 in addition to the most recent 3.x release Actions makes available.

Usage

Once installed and authenticated via repository secrets, schedule the automation from a GitHub Actions workflow (see docs/systemdesign/module_reference.md for the architectural wiring). The CLI remains available for local smoke-tests and packaging checks while the richer workflow management helpers iterate:

keep_github_workflows_active info
keep_github_workflows_active hello
keep_github_workflows_active fail
keep_github_workflows_active --traceback fail
keep_github_workflows_active enable-all-workflows [--owner <user>] [--token <pat>]
keep_github_workflows_active delete-old-workflow-runs [--owner <user>] [--token <pat>] [--keep <n>]
keep-github-workflows-active info
python -m keep_github_workflows_active info

The workflow maintenance commands accept optional flags. When --owner or --token is omitted the CLI falls back to the values discovered via environment variables or the project's .env file. --keep defaults to 50 workflow runs per repository.

Programmatic access is also exposed via the public helpers:

import keep_github_workflows_active as kgwa

kgwa.emit_greeting()
try:
    kgwa.raise_intentional_failure()
except RuntimeError as exc:
    print(f"caught expected failure: {exc}")

kgwa.print_info()

Security

All logging operations sanitize sensitive data before output to prevent credential leakage. The automation redacts:

• GitHub personal access tokens (ghp_, gho_, etc.)
• Authorization headers and API keys
• Token-like patterns in error messages

For complete security documentation including incident response procedures, see SECURITY.md.

Further documentation

• Install Guide
• Development Handbook
• Contributor Guide
• Changelog
• Module Reference
• Security Documentation
• License