accelerate

Run PyTorch training across GPUs with minimal changes.

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./optional-skills/mlops/accelerate/SKILL.md

Quality

Content

86%Weight 40%Scale 1-5

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

The body is a strong, executable reference: concrete code and launch commands cover the common distributed-training cases, and advanced topics are cleanly offloaded to three real reference files. Minor conciseness and explicit-validation-checkpoint gaps keep two dimensions at 4.

Dimension	Reasoning	Score
Conciseness	The body is code- and command-heavy and assumes Claude's familiarity with PyTorch, with only minor padding such as 'Everything else is automatic!' and 'Same code as before!'. It is efficient with a few instances that could be trimmed, fitting the score-4 anchor rather than the fully lean score-5 anchor.	4 / 5
Actionability	Provides copy-paste-ready, executable code and launch commands spanning the common cases (DDP, multi-GPU, multi-node, mixed precision, DeepSpeed, FSDP, gradient accumulation). This matches 'Fully executable; copy-paste ready code or commands; specific examples cover the common cases'.	5 / 5
Workflow Clarity	Workflows 1-5 are clearly sequenced with concrete commands, and the 'Common issues' section provides error-recovery guidance. It is not a 5 because validation checkpoints are not woven explicitly into the workflow steps; it is not a 3 because the sequence and recovery guidance are solid.	4 / 5
Progressive Disclosure	SKILL.md gives a concise overview plus core workflows inline, with advanced topics pushed to three real, one-level-deep reference files (megatron-integration.md, custom-plugins.md, performance.md) that are clearly signaled with descriptive links. This matches 'Clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation'.	5 / 5
	Total	18 / 20 Passed

Description

53%Weight 40%Scale 1-5

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

The description clearly conveys what the skill does but omits any 'when to use' trigger guidance, which caps completeness. Trigger term coverage is adequate but lacks common synonyms like 'distributed' or 'multi-GPU'.

Suggestions

Add a 'Use when...' clause with concrete triggers, e.g. 'Use when distributing PyTorch training across multiple GPUs or nodes, or when the user mentions DDP, FSDP, DeepSpeed, or multi-GPU training.'

Broaden trigger terms to include natural synonyms such as 'distributed training', 'multi-GPU', and 'multi-node' that users actually say.

Optionally enumerate the supported strategies (DDP, DeepSpeed, FSDP, Megatron) to sharpen distinctiveness from a generic PyTorch skill.

Dimension	Reasoning	Score
Specificity	Names the domain (PyTorch training across GPUs) and one concrete action (run training across GPUs) with the 'minimal changes' qualifier, but coverage is not comprehensive. It fits the anchor 'Names domain and 1-2 concrete actions' better than the score-2 anchor whose actions are purely generic.	3 / 5
Completeness	Clearly states what the skill does ('Run PyTorch training across GPUs with minimal changes') but provides no 'Use when...' clause or explicit trigger guidance, so completeness is capped at 3 per the rubric guidelines. It is not a 4 because no 'when' guidance is present even implicitly.	3 / 5
Trigger Term Quality	Includes relevant keywords ('PyTorch training', 'GPUs') but misses common natural variations a user would say such as 'distributed training', 'multi-GPU', or 'DDP'. This matches 'Some relevant keywords but missing common variations or synonyms' rather than the broader coverage of a 4.	3 / 5
Distinctiveness Conflict Risk	The distributed-PyTorch-training niche is mostly distinct with only minor overlap risk against a generic PyTorch or DeepSpeed skill. It is not a 5 because the absence of explicit triggers keeps it from being unambiguously distinguishable.	4 / 5
	Total	13 / 20 Passed

Validation

81%

Warnings & errors only

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

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
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
Path: optional-skills/mlops/accelerate/SKILL.md
Commit: 6680afb

Reviewed: about 3 hours ago

Table of Contents

Discovery Implementation Validation

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.