model-checkpoint-manager

Model Checkpoint Manager - Auto-activating skill for ML Training. Triggers on: model checkpoint manager, model checkpoint manager Part of the ML Training skill category.

1.00x

Quality

Does it follow best practices?

Impact

96%

1.00x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./planned-skills/generated/07-ml-training/model-checkpoint-manager/SKILL.md

Evaluation results

100%

Resumable Model Training Script

Resumable training with complete checkpoint state

Criteria

Without context

With context

Uses PyTorch

100%

Saves model state_dict

100%

Saves optimizer state

100%

Saves epoch number

100%

Saves loss or metric

100%

Loads all saved state

100%

Resumes correct epoch

100%

Configurable interval

100%

No large files

100%

Python script runnable

100%

Without context: $0.2642 · 3m 8s · 15 turns · 14 in / 4,449 out tokens

With context: $0.5375 · 4m 22s · 32 turns · 318 in / 7,234 out tokens

90%

Checkpoint Manager with Disk Space Budget

Best-N checkpoint rotation and cleanup

Criteria

Without context

With context

Python implementation

100%

Configurable keep_top_k

100%

Metric-based ranking

100%

Deletes excess checkpoints

100%

Tracks file paths with metrics

100%

Identifies best checkpoint

100%

Uses ML framework serialization

Demo shows deletion log

100%

No large files produced

100%

Modular design

100%

Without context: $0.2431 · 1m 6s · 15 turns · 16 in / 3,519 out tokens

With context: $0.4335 · 1m 44s · 24 turns · 104 in / 5,555 out tokens

100%

ML Model Registry for Tabular Data Pipeline

Sklearn checkpoint with hyperparameter and experiment metadata

Criteria

Without context

With context

Uses scikit-learn

100%

Framework-native serialization

90%

100%

Hyperparameters saved

100%

Validation metric saved

100%

Timestamp saved

100%

Feature names saved

100%

List all models

100%

Best model retrieval

100%

Loads and verifies

100%

Synthetic dataset only

100%

Without context: $0.2560 · 1m 18s · 17 turns · 16 in / 4,572 out tokens

With context: $0.5195 · 1m 57s · 30 turns · 60 in / 6,847 out tokens

Repository: jeremylongshore/claude-code-plugins-plus-skills
Commit: f17dd51

Evaluated: 18 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Resumable Model Training Script Checkpoint Manager with Disk Space Budget ML Model Registry for Tabular Data Pipeline

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.