Skip to content

manage-deployments.md

Latest commit

 

History

History
79 lines (63 loc) · 3.22 KB

manage-deployments.md

File metadata and controls

79 lines (63 loc) · 3.22 KB
Raw

Manage deployments

A deployment is a declared resource that produces runs. It bundles an agent with runtime bindings, initial events, and an optional schedule.

Declare a deployment

deployments:
  daily-report:
    agent: reporter
    description: "Scheduled daily report (server-side cron + outcome grading)"
    schedule:
      expression: "0 9 * * *"
      timezone: UTC
    initial_events:
      - type: user.message
        content: "Summarize yesterday's commits and generate the daily report."
      - type: user.define_outcome
        description: "Daily report quality gate"
        rubric: "Must include an executive summary and at least three key metrics."
        max_iterations: 3
    resources:
      - type: file
        source: ./data/report-template.md
        mount_path: /data/report-template.md
Field Required Description
agent yes Agent name to run.
agent_version no Pin a versioned agent.
environment no Override the agent's environment.
vaults no Vault bindings.
memory_stores no Memory store bindings.
resources no file, memory_store, or github_repository resources to mount.
initial_events yes 1–50 events that seed each run.
schedule no 5-field cron expression + timezone.

Initial event types

Type Purpose
user.message The user prompt that starts the run.
system.message A system instruction prepended to the run.
user.define_outcome An outcome rubric the run is graded against (description, rubric/rubric_file, max_iterations 1–20).

Resource mounts

Resource type Key fields
file source or file_id, mount_path
memory_store memory_store, access (read_write | read_only), instructions
github_repository url, checkout (branch/commit), mount_path, authorization_token

A file with a local source is uploaded during apply (Files API) and the returned file_id is stored in state; at run time it is mounted at mount_path.

Run a deployment

agents deployment list                  # deployments tracked in state
agents deployment get <name>            # status + resolved bindings
agents deployment run <name>            # trigger a run

Native vs. emulated

Provider Deployment tier What deployment run does
Claude native schedules server-side through the deployments API
Qoder native creates a deployment run and associated session
Bailian, Ark emulated expands into a one-shot session at run time

On the emulated providers, scheduling and outcome rubrics are not enforced server-side — use external cron/CI for always-on or scheduled runs.

Examples