using-claude-cli
Internal contract for invoking the Claude Code CLI non-interactively. Use only when another claude-in-codex skill or a Claude proxy subagent needs to run claude -p, including to support a review workflow, delegate work, produce structured output, or resume or continue a prior Claude CLI session with a follow-up. Do not use as the direct entry point for a user's request to have Claude or Claude Code review code, current changes or the working tree, a branch, or a pull request, even when the request names a specific Claude model; route that request through the sibling adversarial-review skill first. Do not use for conceptual questions about Claude Code or Claude models, including pricing, capabilities, and model comparisons.
- License
- MIT
- Compatibility
- Requires Claude Code CLI on PATH, authenticated and able to run non-interactively with network access.
Pinned to revision 21f357d9389c, so it is the text this page describes rather than whatever the author pushed since.
Files
- skills/using-claude-cli/SKILL.md
- skills/using-claude-cli/agents/openai.yaml
- skills/using-claude-cli/evals/triggers.yaml
- skills/using-claude-cli/references/claude-agent.toml
Every link opens the file at its source, pinned to the revision this page describes.