run

Run a Cortivex pipeline by name or generate one from a description

Run a Cortivex AI agent pipeline. You can specify a pipeline name (from templates or saved pipelines) or describe what you want in natural language.

Usage

Run a named pipeline:

/run pr-review
/run security-audit
/run full-test-suite

Run with natural language:

/run Review this PR for security issues and generate tests
/run Scan for vulnerabilities and fix them
/run Refactor the auth module and update docs

What happens:

  1. Cortivex loads the pipeline definition (YAML) or generates one from your description
  2. The DAG executor resolves node dependencies and execution order
  3. Each node spawns a specialized Claude agent with expert system prompts
  4. Agents coordinate via mesh protocol to prevent file conflicts
  5. Results stream in real-time via WebSocket to the dashboard
  6. Execution history is recorded for the learning engine to optimize future runs

Available templates:

  • pr-review, security-audit, full-test-suite, js-to-typescript
  • bug-hunt, refactor-module, performance-audit, api-design
  • coverage-boost, documentation-refresh, onboarding-guide
  • nightly-review, dependency-update, changelog-release, pre-release-check