# Central Agentic Ops agent resources

> Machine-readable navigation for the authoritative CAO documentation and dashboard architecture.

Use these routes to navigate the documentation and dashboard architecture without executing client-side JavaScript. Each HTML documentation page advertises a compact JSON representation containing stable identity, relationships, provenance, and freshness.

- [Complete machine-readable resource index](https://githubnext.github.io/gh-cao/agent/resources.json)

## Important resources

- [Agent analysis](https://githubnext.github.io/gh-cao/agent-analysis/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/agent-analysis/index.json)
  - Give an agent the dashboard's own pages and named queries through gh cao or a read-only MCP server, with or without shell access.
- [At a glance](https://githubnext.github.io/gh-cao/dashboard/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/dashboard/index.json)
  - See what is moving, what needs attention, and whether the evidence is ready for follow-up.
- [Build Your First Campaign](https://githubnext.github.io/gh-cao/author-your-first-operation/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/author-your-first-operation/index.json)
  - Define one repository outcome, generate the workflows, and prove the campaign in review mode.
- [CAO Activity](https://githubnext.github.io/gh-cao/activity/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/activity/index.json)
  - Learn how CAO Activity collects gh-aw logs for Central Agentic Ops.
- [CAO Commands](https://githubnext.github.io/gh-cao/cao-cli/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/cao-cli/index.json)
  - Configure and inspect CAO with the Go extension, and prepare evidence with the repository-local pipeline CLI.
- [Central Agentic Ops (CAO)](https://githubnext.github.io/gh-cao/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/index.json)
  - Monitor GitHub Agentic Workflow runs, costs, and outcomes. Control execution scope and rollout policy.
- [Dashboard Language](https://githubnext.github.io/gh-cao/dashboard-language/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/dashboard-language/index.json)
  - Build dashboard views from trusted data with readable, declarative queries.
- [Data model](https://githubnext.github.io/gh-cao/dashboard-data-model/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/dashboard-data-model/index.json)
  - Understand the canonical entities, relationships, identities, and lifecycle of Central Agentic Ops dashboard data.
- [How the Control Plane Works](https://githubnext.github.io/gh-cao/architecture/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/architecture/index.json)
  - Understand the control plane's purpose, execution boundary, and core safety properties.
- [Intelligence](https://githubnext.github.io/gh-cao/intelligence/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/intelligence/index.json)
  - Understand how CAO turns canonical operational evidence into deterministic, review-only Decisions.
- [Local Worker Debugging](https://githubnext.github.io/gh-cao/local-debugging/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/local-debugging/index.json)
  - Use a local debugger to launch one CAO worker, inspect staged proposals, and iterate without a campaign dispatcher.
- [Monitor, Recover, and Maintain](https://githubnext.github.io/gh-cao/operations/)
  - [Machine-readable metadata](https://githubnext.github.io/gh-cao/operations/index.json)
  - Monitor control-plane runs, stop unsafe activity, recover from incidents, and maintain installed campaigns.

## Deeper access

- [Dashboard agent access guide](https://githubnext.github.io/gh-cao/cao/llms.txt): choose CAO skills, bounded Activity artifacts, targeted MCP queries, or bulk SQLite analysis without scraping the dashboard UI.
- [Agent analysis](https://githubnext.github.io/gh-cao/agent-analysis/): use the CAO CLI or read-only MCP server for large operational datasets and named Dashboard Language queries.
- [Complete documentation](https://githubnext.github.io/gh-cao/llms-full.txt): use only when the compact routes do not answer the task.