mlops-training-triage

A process for investigating failed or costly machine-learning model training jobs. It checks the training job, platform health, and a safe retry or checkpoint plan.

In plain words
What is it for?
Use it to inspect failures, resource use, queue time, checkpoints, and runtime costs, then plan a safe retry.
Why use it?
It helps separate problems in the code or training data from problems in the computing platform. This avoids unsafe retries and helps preserve useful checkpoints.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/agenticdevops/aoh/mlops-training-triage
Any agent
npx skills add agenticdevops/aoh --skill mlops-training-triage
Clone the repo
git clone --depth 1 https://github.com/agenticdevops/aoh

Made for: Claude Code, Codex.

Per session 36 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 123 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00036 $0.00123
Opus 5 $0.00018 $0.00062
Sonnet 5 $0.00007 $0.00025
Haiku 4.5 $0.00004 $0.00012

Measured 2d ago against content hash f6cc1a2bd9db, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mlops-training-triage scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

examples/acme-platform-ops/skills/mlops-training-triage/SKILL.md · 16 lines

What it actually says

MLOps Training Triage

Process skill: triage failed or expensive model training jobs.

Process

  1. Use the ml-training-job-triage skill to inspect job failures and utilization signals.
  2. Use the service-health-report skill to rule out platform-level causes.
  3. Separate data/code issues from infrastructure issues.
  4. Recommend a safe retry and checkpoint strategy.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 16 lines · 36 tokens per session scan A f6cc1a2bd9db

Subscribe to this mod's changes

mlops-training-triage is a skill published in the GitHub repository agenticdevops/aoh (5 stars, last pushed 1mo ago), licensed MIT. It adds 36 tokens to every session and 123 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.