What is analyze-results?

A skill for comparing machine-learning training runs or saved checkpoints and choosing the best result. A checkpoint is a saved version of a model during training.

It is for ranking runs, selecting a checkpoint to keep or deploy, and identifying untested checkpoints that should be evaluated.

It removes the need to compare experiment results manually and makes the meaning of “best” explicit through a chosen primary measurement and direction.

How EricYuan2007/GPUPlane configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to EricYuan2007/GPUPlane. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/EricYuan2007/GPUPlane/main/.claude/skills/gpu-training/analyze-results/SKILL.md

Clone the repo

git clone --depth 1 https://github.com/EricYuan2007/GPUPlane

Made for: Claude Code.

Or install gpu-training, the plugin that ships this one along with the rest of its 3 skills, 1 MCP server.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/ericyuan2007/gpuplane/analyze-results.svg)](https://agentmods.dev/skills/ericyuan2007/gpuplane/analyze-results)

Your own site

<a href="https://agentmods.dev/skills/ericyuan2007/gpuplane/analyze-results"><img src="https://agentmods.dev/badge/skills/ericyuan2007/gpuplane/analyze-results.svg" alt="Measured on agentmods" height="20"></a>

Per session 47 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 968 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00047	$0.00968
Opus 5	$0.00023	$0.00484
Sonnet 5	$0.00009	$0.00194
Haiku 4.5	$0.00005	$0.00097

Measured 7d ago against content hash 342a47fe2198, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

analyze-results 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 7d 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.

.claude/skills/gpu-training/analyze-results/SKILL.md · 84 lines

How it starts

The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.

analyze-results

Pick the best run / checkpoint from a batch of experiments. The control plane already resolves the primary metric through experiment → project → global inheritance and knows its direction — compare_* and get_best_checkpoint rank by it. Call list_metric_definitions with the relevant scope to confirm the effective choice. If none is set, use set_primary_metric after agreeing on the decision scalar (otherwise "best" is undefined).

Workflow

Which runs to compare? Gather the run ids from the user, or list_jobs(status=SUCCEEDED) → their run ids. For a single experiment, get_best_checkpoint(experiment_id=…) gives the answer directly.
Compare runs — compare_runs(ids=[…]) returns a ranked table + a best_run_id. Read direction (maximize/minimize) so you describe the ranking correctly to the user.
Compare checkpoints — compare_checkpoints(ids=[…]) ranks a run's checkpoints by their evaluation results (not training metrics). Checkpoints without evaluations show primary_value=null — they're untested, not worse. The next_actions field tells you which to evaluate. If evaluations use different suite/dataset/version values, comparison stops; pass the same suite, dataset, and dataset_version filters instead of ranking incomparable evidence.
Confirm the best — get_best_checkpoint(run_id=…) (or experiment_id=…) returns the single recommended checkpoint, preferring evaluation results and falling back to training-run best values.
Fill evaluation gaps — if the best run's checkpoints are unevaluated, evaluate_checkpoint(checkpoint_id=…, command=[…]) queues an EVALUATE job (inherits the training job's working_dir/resources). Wait via monitor-experiment, then re-compare.

Keep / revert loop (Karpathy autoresearch pattern)

The point of comparing is to decide what to keep. After picking the best checkpoint:

Keep: tell the user the checkpoint path (from compare_checkpoints entries' .path) to commit / tag / deploy.
Revert: if a run lost, tell the user to git revert the config change that produced it and retry. The control plane records the run's config + source so you can attribute each result to a change.

Read the full file on GitHub · 84 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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.

7d ago First seen · 84 lines · 47 tokens per session scan A 342a47fe2198

Subscribe to this mod's changes

analyze-results is a skill published in the GitHub repository EricYuan2007/GPUPlane (0 stars, last pushed 13d ago), licensed Apache-2.0. It adds 47 tokens to every session and 968 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.

Analytics and metrics Data and AI

agent-platform-rag-engine-management

Manage and query Agent Platform RAG Engine Corpora and retrieve grounded contexts using the Google GenAI SDK. Use when listing RAG corpora or files, inspecting a corpus, retrieving contexts, or generating content grounded in a RAG corpus. Do not use for standard database queries (use SQL/Spanner skills), Google…

google/skills · 85 tokens

agent-platform-model-registry

Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.

google/skills · 60 tokens

foundry-config-setup

Resolve missing setup caused by a hardcoded Foundry project endpoint or model in a sample. Use when a sample fails because it uses a placeholder/hardcoded projectendpoint (for example "https://your-project.services.ai.azure.com") or a hardcoded model instead of reading them from the environment.

microsoft/agent-framework · 65 tokens

google-cloud-solution-agentic-analytics-spark-knowledge-catalog

Discovers requirements and generates guidance to design and deploy a governed, secure agentic-analytics solution for data that's distributed across Google Cloud, other cloud providers, or on-premises. Data that's outside Google Cloud (such as data from Databricks, Snowflake, Salesforce, SAP, or Oracle systems) is…

google/skills · 138 tokens

training-check

Interactively monitor training metrics from the current Codex session, periodically checking WandB or fallback logs for NaN, divergence, plateaus, and broken runs.

wanshuiyin/Auto-claude-code-research-in-sleep · 35 tokens

nemo-automodel-launcher-config

Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.

NVIDIA/skills · 30 tokens

GPUPlane: Skill for Claude Code