Skip to content

Ways to Use Engineering Runtime

The same runtime binary serves three consumers. Behaviour is identical — Bootstrap → Context → Policy → Auth → Execution → Audit — only the caller changes.

┌─────────────┐   ┌─────────────┐   ┌──────────────────┐
│   Human     │   │   CI/CD     │   │  AI / Runtime    │
│  engineer   │   │  pipeline   │   │     Agent        │
└──────┬──────┘   └──────┬──────┘   └────────┬─────────┘
       │                 │                   │
       └────────────┬────┴───────────────────┘
                    ▼
              runtime binary
                    ▼
           Engineering platforms
Mode Who drives Typical entry Guide
Local (human) You, in a terminal runtime github …, runtime capability execute … Local setup
CI/CD A pipeline job Composite action → runtime capability execute … CI/CD
AI agent Claude / Cursor / any agent under the Runtime Agent contract Agent may only invoke runtime … AI agent

What every mode shares

  1. One binary. Install from engineering-runtime-releases (or build from source). No daemon, no language runtime.
  2. One Runtime Home for contracts and working state. Specs, commands and audit live there. Config/policy, capability definitions, tool context and provider credentials keep their external owners — see Runtime Home.
  3. Two execution surfaces.
  4. Provider operations: runtime github …, runtime files …
  5. Command Engine: runtime command run <binary> …
  6. Capabilities. Markdown workflows that compose the surfaces above — runtime capability validate|execute ….
  7. Governance before work. Policy can deny before auth or execution; every outcome (including denial) is audited.

Pick a path

  1. Install
  2. Quick Start
  3. Deeper local notes: Local setup
  1. Add the setup-runtime action
  2. Set RUNTIME_GITHUB_TOKEN only for GitHub work; pass any org explicitly
  3. Call runtime capability execute … or provider commands
  4. Full contract: CI/CD
  1. Install runtime and bootstrap
  2. Open the engineering-runtime-agent repo (or apply the same hooks/rules)
  3. Ask in natural language — the agent may only call runtime
  4. Full contract: AI agent

Identity on the audit trail

Set RUNTIME_CONSUMER for the caller class, plus optional actor/session labels so audit records show which run drove the binary:

Value Use when
human Interactive terminal (default when unset and not CI)
ci Pipelines — samples set this explicitly
ai Runtime Agent sessions 
export RUNTIME_CONSUMER=ai
export RUNTIME_ACTOR_NAME=repo-maintenance-agent
export RUNTIME_SESSION_ID=task-123
runtime github user get
runtime audit tail -n 1
Repository Role
engineering-runtime-releases Versioned binaries + checksums
engineering-runtime-capabilities Shared / company capability store
engineering-runtime-ci Working CI workflows + setup-runtime action
engineering-runtime-agent Runtime Agent contract + shell hooks
Engineering Runtime site Vision and architecture essays

Operational examples on this site were verified against Runtime 0.9.8. After bootstrap, the version-exact files in Runtime Home win.