agentsclimarketplace

Super ai ml ops

Skill arpitexplores/skills-super/super-ai-ml-ops

SUPER Skills catalogue: portable Markdown agent skills for AI vibe coding, SEO, AI SEO/GEO, marketing, design, AI agents, DevOps, security, automation, data, business, and operations.

Install
npx -y skills add arpitexplores/skills-super --skill super-ai-ml-ops

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

AI/ML operations: evaluation, monitoring, cost control, and reliability. Use for productionizing AI systems.

SKILL.md

1.6 KB, as published. Nobody here has run it

Super AI/ML Ops

Overview

Make AI systems stable and measurable in production.

User Intent Examples

  • "Need help with LLM Evaluation for my product/site."
  • "Create a plan for LLM Ops."
  • "Not sure where to start, need a quick assessment."

Workflow

  1. Define evaluation metrics, datasets, and acceptance thresholds.
  2. Set up observability for quality, latency, and errors.
  3. Implement caching and cost controls.
  4. Create monitoring and alerting for regressions.
  5. Establish release and rollback procedures for prompts/models.
  6. Document runbooks and ongoing QA cadence.

Minimal Intake Questions

  • Primary goal or outcome
  • Scope (pages, systems, teams, or timeframe)
  • Constraints (tools, budget, timeline)

Output Format

  • Eval plan and scoring rubric
  • Monitoring and alerting checklist
  • Cost and caching strategy
  • Release and rollback plan
  • Runbook and QA cadence

Routing Map (Modules)

  • LLM Evaluation -> references/modules/llm-evaluation.md
  • LLM Ops -> references/modules/llm-ops.md

Bundled References

  • references/modules/
  • scripts/
  • assets/
  • agents/

Compatibility Notes

  • If any module references slash commands or tool-specific paths, translate them into plain-language steps.
  • Keep outputs platform-agnostic unless the user specifies a specific tool, stack, or agent.

Guardrails

  • Do not rely on single metrics; include qualitative checks.
  • Track cost per request and cap budgets.
  • Treat prompt/model updates as production changes.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.