This page routes humans and agents to the right AgentFlow SDLC document without making the README carry every detail.
- Read
agentflow-in-5-minutes.md. - Use
get-started.mdfor the onboarding path. - Review examples in
examples/.
- Start with
assisted-onboarding.md. - Check environment expectations in
environment-tools.md. - Use
project-setup.mdandproject-config.mdto choose branch, validation, and routing defaults. - Use
assisted-update.mdfor existing installations.
- Read
../AGENTS.md. - Read the active adapter file for your executor.
- Read
agent-workflow.md. - Read
issue-standards.md. - Read the active issue or
SPEC.md. - Use
guides/contribution-workflow.md.
| Concept | Doc |
|---|---|
| Product overview | agentflow-in-5-minutes.md |
| Workflow phases | agent-workflow.md |
| Intelligent collaboration | intelligent-collaboration.md |
| Execution targets | execution-targets.md |
| Portable capabilities | capabilities.md |
| Project config | project-config.md |
| Release versioning | release-versioning.md |
| Goal Command Center | cockpit.md |
Use this deterministic entry sequence:
AGENTS.md- active adapter (
CLAUDE.md,CODEX.md,AGY.md, or equivalent) docs/agent-workflow.mddocs/issue-standards.md- active issue or
SPEC.md
Then use role packages under agents/roles/ and workflow skills under agents/workflows/ as needed.
Use Cockpit as optional Goal Command Center:
AGENTFLOW_REPOSITORIES=owner/repo agentflow-sdlc cockpit
agentflow-sdlc cockpit doctor --jsonCockpit is first-class in product docs and release gates, but runtime opt-in. Normal SDLC work does not require starting a service, opening a port, or configuring OAuth.