Automate

Agents

Goals, schedules, memory grants and run history.

Availability note — these docs describe both working and planned product areas. Account authentication, administrator-managed AIHK/AICA tenant regions and keys, signed-in Composer runs and public Guard scans are connected. Skills, Memory, Agents, Workflows, Apps, billing, the public SDK/CLI and the unified API are previews unless a section explicitly says otherwise.


Roadmap and package-contract reference: this page describes the intended GenCMS runtime. The current deployment persists workspace records and drafts but does not execute marketplace Skills, index Memory for retrieval, publish packages, or run Agents, Workflows or Apps. See the deployed API reference for the routes available today.

An Agent is a persistent worker: a model, the Memory it may read, the Skills it may call, a goal and a trigger. It keeps state between runs, which is what separates it from a Workflow you happen to run twice.

agent.yaml
name: Weekly SEO Monitor
goal: Monitor selected webpages and report what changed.

model:
  mode: auto

memory:
  read:
    - website-knowledge
    - brand-rules
  write:
    - project-state

skills:
  - gencms/[email protected]
  - gencms/[email protected]

schedule:
  cron: "0 8 * * 1"
  timezone: Asia/Hong_Kong

output:
  - email
  - workspace-report

Triggers

  • Schedule — cron, with a timezone
  • Event — a Memory source changed, a Workflow completed, a Guard scan crossed a threshold
  • Webhook — an external system calls the Agent's endpoint
  • Manual — a run started from the workspace or the API

Run history

Every run records its trigger, the model chosen, the Skills called, the Memory retrieved, credits spent, duration and outcome. An Agent you cannot audit is an Agent you will eventually switch off.