Skip to main content

Agents

Autonomous services that extend your hub with background processing, domain intelligence, health monitoring, and more. Organized under domain controllers or as shared infrastructure.

What Is an Agent?

An Agent is an independent service that plugs into a hub via a standardised protocol. Agents handle work that doesn't belong in a request/response cycle — SEO analysis, content quality auditing, health monitoring, background jobs, scheduled tasks, event-driven workflows, and third-party integrations.

Agents come in two kinds: work agents operate under a domain controller's direction, and infrastructure agents provide shared system services used by any domain.

Autonomous

Agents run independently of your request handlers. They start, stop, and scale on their own. Crash one and the rest keep running.

Protocol-Based

Agents communicate via a defined protocol — message types, lifecycle events, health checks. Implement the protocol in any language; connect to any Weblisk server.

Blueprint-Powered

Agents are defined using Blueprints. A blueprint declares the agent's capabilities, inputs, outputs, and triggers — the implementation just fulfils the contract.

Sandboxed

Each agent runs in its own context with explicit capability grants. An agent can only access the resources it's been given — no ambient authority.

Agent Catalog

Weblisk ships with three domain agent families and four infrastructure agents:

Agent Kind Description Tier
seo-analyzer Work (SEO) Title tags, meta descriptions, heading hierarchy, structured data Free
a11y-checker Work (SEO) WCAG compliance, color contrast, ARIA roles, focus management Free
content-analyzer Work (Content) Readability scoring, heading structure, paragraph balance Free
meta-checker Work (Content) Open Graph, Twitter cards, JSON-LD, canonical URL validation Free
uptime-checker Work (Health) Endpoint availability, response time, TLS certificate grading Free
perf-auditor Work (Health) Core Web Vitals, resource budgets, load performance Free
sync Infrastructure Background data synchronisation between client store and server Free
cron Infrastructure Scheduled task execution with cron-style expressions Free
webhook Infrastructure Inbound and outbound webhook processing with retries Free
email-send Infrastructure Transactional email sending with templates and queuing Free
ai-agent Work LLM integration with prompt management, tool use, and streaming Pro
search-agent Work Full-text and semantic search indexing with incremental updates Pro
media-agent Work Image/video processing, thumbnails, and format conversion Pro

How Agents Work

agents/sync.yaml
name: sync
version: 1.0.0
description: Syncs client-side IDB store with server database

triggers:
  - event: client.sync.push
  - schedule: "*/5 * * * *"    # every 5 minutes

capabilities:
  - database:read
  - database:write
  - realtime:publish

config:
  conflict_resolution: last-write-wins
  batch_size: 100
Browse all available agents in the Agent Catalog, or learn how to build your own in the Building Agents guide.