runshift / docs
docs
the control plane for AI agents. these docs cover how to run it.
what is runshift
runshift is the control plane for builders running multiple AI agents simultaneously. scan your repo, coordinate agents working in parallel, decide what matters.
it's model-agnostic. Claude Code, Cursor, or any agent connects to the same coordination layer — file locking, session management, and trust gates across all of them.
how it works
scan
the entry point. run npx runshift init in any repo. runshift maps the codebase, flags security gaps and blast-radius files, and creates coordination rules for files, sessions, and risky actions.
coordinate
runshift keeps Claude Code, Cursor, and external agents aligned around the same repo context, active work, and approval gates. connect via @runshift/connect — MCP execution hooks, file locking, and session management.
decide
trust gates fire before consequential actions. you approve or deny from the dashboard or Slack. every decision is logged in a full audit trail.
AMP
the Agent Message Protocol. the reporting contract between agents and runshift. any agent — Claude, GPT, LangChain, custom — signals runshift before acting. full spec at runshift.ai/amp.
in these docs
getting started
how to set up runshift, run your first agent, and read the deck.
the deck
outcome cards, agent status, cost tracking, and how to read the execution view.
gates
what trust gates are, how they fire, and how to act on them from Slack or the dashboard.
AMP spec
the Agent Message Protocol. the reporting contract between agents and runshift.
claude code
connect claude code to runshift. every tool call signals your control plane.