Red Hat AI AgentOps
Connect any agent framework to the Red Hat OpenShift AI platform.
AgentOps provides documentation and working code examples for connecting agent frameworks to Red Hat OpenShift AI. Every page on this site is designed to be read by humans in a browser and consumed by AI agents as structured markdown — the same content, two formats.
Topics
- Basic Agents — Build your first agent with step-by-step examples across popular frameworks.
- Security — Sandbox code execution, authenticate workloads, and enforce policy for agents.
- Tracing — Send agent traces from any framework to MLflow on OpenShift AI.
- Evaluation — Measure agent quality, correctness, and safety.
- Identity — Workload identity for agents using SPIRE, SPIFFE, and service accounts.
- Observability — Metrics, logs, and dashboards for agent operations.
- Catalog — Discover, register, and version agents, tools, and prompts.
- Lifecycle — Build, deploy, update, and retire agents on OpenShift.
Built for Three Audiences
This site serves the same content in three ways. Humans get a rich browsable UI with syntax-highlighted code, tabbed examples, and responsive navigation. AI agents get clean markdown with the same content — no scraping, no parsing HTML, no losing context to boilerplate. And Claude Code gets a plugin marketplace — skills can be installed directly into a session with a single command.
| Audience | Access | Format |
|---|---|---|
| Humans | Browse the site | HTML with full UI |
| AI agents | /AGENTS.md or append .md to any page URL | Plain markdown |
| Claude Code | /plugin marketplace add | Installable skills |
Point your agent here
Add the site URL to your agent's context so it can fetch documentation on demand. In Claude Code, add a reference to your CLAUDE.md:
## Reference Documentation
- AgentOps docs: https://agentops.redhatskills.com/AGENTS.mdFor any AI coding agent or LLM tool, point it at the AGENTS.md URL — it lists every page with its markdown URL:
https://agentops.redhatskills.com/AGENTS.mdYour agent can fetch that index, find the relevant page, and retrieve the full markdown content — no API key required, no rate limits.
AGENTS.md
The AGENTS.md file at the site root is an index of all documentation pages with their markdown URLs. It follows the emerging agentskills.io convention — a machine-readable entry point that tells agents what content is available and where to find it.
| URL | Format | Audience |
|---|---|---|
/AGENTS.md | Markdown | AI agents — page index with all markdown URLs |
/tracing/connect-to-mlflow | HTML | Humans — rich UI with syntax highlighting and navigation |
/tracing/connect-to-mlflow.md | Markdown | AI agents — same content as plain markdown |
View as Markdown
Every content page includes a “View as Markdown” link at the top. Click it to see the raw .md source that agents consume. Append .md to any page URL to get the markdown version directly:
# HTML (for humans)
https://agentops.redhatskills.com/identity/using-spiffe-spire
# Markdown (for agents)
https://agentops.redhatskills.com/identity/using-spiffe-spire.mdSkills Marketplace
This site is also a Claude Code plugin marketplace. Skills are reusable agent capabilities — scaffolding generators, code reviewers, deployment helpers — that you can install directly into your Claude Code session.
Add the marketplace to Claude Code:
/plugin marketplace add eformat/agentops-redhatskills-comThen install individual skills:
/plugin install langchain-agent@agentops-redhatskills| Skill | Description |
|---|---|
langchain-agent | Scaffold a LangChain/LangGraph ReAct agent |
crewai-agent | Scaffold a CrewAI agent with role-based orchestration |
autogen-agent | Scaffold an AutoGen multi-agent conversation |
llamaindex-agent | Scaffold a LlamaIndex ReAct agent |
google-adk-agent | Scaffold a Google ADK agent with session management |
agent-deploy-openshift | Deploy a scaffolded agent to OpenShift using UBI 9 S2I |
mlflow-tracing | Add MLflow tracing to a scaffolded agent |
agent-sandboxing | Add a secure code sandbox with defense-in-depth isolation |