CtrlK
BlogDocsLog inGet started
Tessl Logo

pytorch-lightning

High-level PyTorch framework with Trainer class, automatic distributed training (DDP/FSDP/DeepSpeed), callbacks system, and minimal boilerplate. Scales from laptop to supercomputer with same code. Use when you want clean training loops with built-in best practices.

64

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./optional-skills/mlops/pytorch-lightning/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

72%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable skill body with executable code for every workflow and clean progressive disclosure to real reference files. It loses points on conciseness from celebratory commentary and on workflow clarity from absent validation checkpoints in training/distributed workflows.

Suggestions

Trim editorial commentary ("That's it!", "automatic!", star/download counts) so every token earns its place and conciseness moves toward the score-3 anchor.

Add explicit validation checkpoints to distributed and training workflows (e.g., verify DDP ranks sync, confirm checkpoints save, check val_loss decreases) to lift workflow clarity above 2.

Replace framework-jargon triggers in the body's "when to use" section with natural phrasings users would actually say (e.g., "train on multiple GPUs", "add early stopping").

DimensionReasoningScore

Conciseness

Code is executable and mostly efficient, but it is padded with commentary like "That's it!", "No .to('cuda') needed!", "automatic!", and marketing fluff ("battle-tested", "⭐ 29,000+"), fitting the mostly-efficient-but-could-be-tightened anchor rather than the lean score-3 example.

2 / 3

Actionability

Provides fully executable, copy-paste-ready code across all five workflows (LightningModule, validation/test steps, DDP, callbacks, LR scheduling), matching the fully-executable anchor rather than the pseudocode score-2 case.

3 / 3

Workflow Clarity

The 3-step conversion and validation/test workflows are clearly sequenced, but training workflows lack explicit validation checkpoints or validate→fix→retry feedback loops for risky operations like distributed runs, fitting the sequence-present-but-checkpoints-missing anchor rather than the score-3 example with explicit validation steps.

2 / 3

Progressive Disclosure

The body is a concise overview with a clearly signaled one-level-deep "Advanced topics" section linking to real bundle files (callbacks.md, distributed.md, hyperparameter-tuning.md), all confirmed present, matching the clear-overview-with-well-signaled-references anchor.

3 / 3

Total

10

/

12

Passed

Description

85%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that names concrete capabilities and includes an explicit "Use when" trigger, distinguishing it clearly from sibling skills. The main weakness is trigger-term phrasing that favors framework jargon over the natural language users would actually say.

DimensionReasoningScore

Specificity

Lists multiple concrete actions: "Trainer class", "automatic distributed training (DDP/FSDP/DeepSpeed)", "callbacks system", and "minimal boilerplate" — matches the anchor for multiple specific concrete actions, not the score-2 anchor of domain-plus-some-actions.

3 / 3

Completeness

Explicitly answers what it does (the capability list) and when to use it ("Use when you want clean training loops with built-in best practices"), matching the clearly-answers-both anchor rather than the score-2 case where "when" is only implied.

3 / 3

Trigger Term Quality

Includes relevant keywords ("training loops", "distributed training", "DDP/FSDP/DeepSpeed", "callbacks") but leans on framework-specific jargon and misses common natural variations a user would say like "train a PyTorch model" or "multi-GPU training", fitting the some-relevant-keywords-but-missing-variations anchor.

2 / 3

Distinctiveness Conflict Risk

Targets a clear niche (PyTorch Lightning, DDP/FSDP/DeepSpeed, callbacks) with distinct triggers unlikely to fire for the wrong skill; uses third-person voice with no penalty.

3 / 3

Total

11

/

12

Passed

Validation

81%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

13

/

16

Passed

Repository
NousResearch/hermes-agent
Reviewed

Table of Contents

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.