/code @devintern/pm DevIntern + Claude Code
Run Claude Code as the agent behind both planning and implementation. DevIntern formats tracker context, invokes the CLI headlessly, and captures commits and PR output on your machine.
How it works
What DevIntern does with Claude Code
Each product path keeps Claude Code as the handoff surface. No rip-and-replace, no new dashboard.
/code Implement tickets with Claude Code
- 1 Install Claude Code and set AGENT_HARNESS=claude-code.
- 2 DevIntern builds a task prompt from tracker data and repo context.
- 3 Claude Code runs non-interactively; changes land in git and optional PRs.
/pm Draft specs with Claude Code
- 1 Same harness drives @devintern/pm story generation.
- 2 Figma MCP support available when Claude Code is configured.
- 3 Output is a reviewable draft before posting to your tracker.
Capabilities
Built for real Claude Code workflows
Default harness in both @devintern/code and @devintern/pm
Headless execution with permission-skipping flags when configured
Model override via configuration
Figma MCP integration for PM flows (Claude Code only)
Runs locally with your Anthropic credentials
Get connected
Setup guides and configuration
Credentials stay in project-local .env files. DevIntern runs on your machine or server, not a shared implementation cloud.
FAQ
Common questions about Claude Code
Is Claude Code required?
No. It is the default, but you can set AGENT_HARNESS to opencode, codex, or cursor without changing the rest of your setup.
Where are API keys stored?
In your agent CLI configuration and project .env files on your machine. DevIntern does not host or proxy model calls.
Try DevIntern with Claude Code
Connect Claude Code, run your first workflow, and keep the trackers, repos, and reviewers your team already trusts.
14-day trial · perpetual license · BYO model keys