Runtime flows

Use these traces to follow an OpenClaw Enterprise request or background operation through the current source. They describe implementation order, state changes, and failures. For supported product behavior, use the feature reference linked from each trace; for commands to run locally, see Testing.

Startup and infrastructure

Console, configuration, and Agents

Platform console, Agent creation, Agent editing, workspace files, and stop and deletion
Agent Presets: template admission, variable rendering, and independent draft creation
Agent provisioning: separate Secret saving, queued setup, resource creation, safe retries, and first deployment handoff
Agent native admin UI: console access, shared-session admission, and private gateway proxying
Agent runtime logs: tiered authorization, view audit, Kubernetes Pod and log reads, and the sanitizer
Configuration Driver and configuration persistence and version admission
Harness execution topology
Workspace files and Agent plugins
Production terminal interface and native client behavior

Identity and credentials

Agent repository access, credential service, and repository configuration
Repository credential tests; the original RFC and qualification record preserve proposal and historical evidence separately from current support
Namespace IAM policy: authorized Role and AccessBinding changes and audit commit
Local password authentication and service API keys
Secret storage and delivery
Credential source lifecycle: gateway registration, Agent binding, admission, and retried deletion
Harness authentication binding and ServiceAccount Driver credential delivery

Drivers and placement

Continuous integration

GitHub Actions testing and test preparation
ClawSweeper dispatch: hosted admission and review handoff
Run and diagnose CI checks