Claude Code Base Action
A low-level GitHub Action wrapper for running Claude Code with caller-defined prompts and permissions.
What it helps with
Claude Code Base Action is the thin lower-level building block behind custom GitHub workflows. It installs and runs Claude Code with the caller's inputs, but its README explicitly says it does not enforce trust boundaries. Running it in a repository is equivalent to running Claude Code there, including reading project-level configuration. Anthropic directs workflows that process issues, fork pull requests, or external comments to the higher-level claude-code-action instead.
Best fit
Best only for security-experienced teams with fully trusted inputs and custom workflow needs; exclude from the general directory in favor of the supported action.
Setup reality
Needs developer/security engineer for workflow YAML, allowed tools, secrets or workload identity, repository trust, and GitHub permissions.
License context
Permissive licenseBusiness use and modification are generally allowed. Keep required notices and review trademarks, dependencies, and hosted-service terms.MIT. Commercial use and modification are permitted with notice retention; Claude Code and provider services remain separately licensed.
Resource facts
- Use case
- AI Agent Tools & UIs
- Industries
- Developers
- Setup
- Needs developer
- Audience
- Technical
- Capabilities
- No catalog chips
Last verified . Source: github.com