Skip to content

0.5. Resources

Which sources should you trust first?

Agent APIs change quickly. Prefer the primary documentation and the locked source used by this repository:

Blog posts are useful for context, but the installed version, API schema, lockfile, and tests decide what this repository actually supports.

Which open-source coding assistants can help?

goose, Gemini CLI, and other open-source coding agents can inspect this repository, run its tasks, and explain unfamiliar code. VSCodium provides a community-built editor distribution when you want an open-source VS Code codebase without Microsoft's branded build.

An assistant is an untrusted collaborator, not an authority. Review patches, protect credentials, and require the same mise run check and mise run test gates as human changes. The repository's AGENTS.md gives compatible tools one shared set of project rules.

What are Agent Skills?

An Agent Skill is a small, discoverable instruction package rooted at SKILL.md. Unlike the always-on repository guidance in AGENTS.md, a skill is loaded only for a relevant task. The Ops Copilot demonstrates least-privilege discovery: it can list available skills and load one allowed skill by name instead of injecting every instruction into every prompt.

You will implement and test that boundary in 3.2. Skills.

Where can you ask questions or contribute?

The course repository is public at MLOps-Courses/agentops-open-course:

  • Use course issues for reproducible defects and scoped improvements.
  • Read CONTRIBUTING.md before opening a pull request.
  • Report vulnerabilities through the private process in SECURITY.md, never a public issue.
  • Join the upstream AAIF, CNCF, ADK, agentgateway, kagent, MLflow, MCP, and A2A communities when a question belongs to the project itself.

Maintainers should still apply the anonymous publication gate in 8.4. Documentation whenever repository, Pages, DNS, or source-link contracts change.

What companion material is available?

The MLOps Coding Course covers broader software and model-lifecycle foundations. In this repository, agents/python is the canonical implementation and infra is the canonical deployment contract; each documentation excerpt links back to those sources.

Where do the key terms live?

The full, alphabetical 0.7. Glossary defines every course term — agent, tool, MCP, A2A, RAG, HITL, guardrail, trace, WIF, and the rest — and links each one back to the section that introduces it. Start there when a term is unfamiliar.

How do you spot a stale resource?

Check its publication date, compare examples to the version pinned in this repository, and verify every field against the current schema. If a course link or command has drifted, open a documentation issue with the page, observed output, and your platform.