CtrlK
BlogDocsLog inGet started
Tessl Logo

data-loading

Optimize data loading pipeline to prevent GPU starvation. Use when setting up DataLoader or data preprocessing.

63

Quality

73%

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 ./researchclaw/skills/builtin/tooling/data-loading/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

72%Weight 40%Scale 1-3

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

The body is concise and well-organized with concrete, directly-applicable settings, but it reads as a tip checklist rather than an executable recipe or sequenced workflow. Adding an assembled code example and a verification step would lift actionability and workflow clarity.

Suggestions

Include one copy-paste-ready DataLoader(...) code block that combines num_workers, pin_memory, and persistent_workers so the guidance is fully executable.

Add a validation/checkpoint step such as profiling with torch.utils.bottleneck before and after changes to confirm the pipeline is no longer GPU-starved.

Reorder the items into a short sequence (configure -> profile -> iterate) so the checklist reads as a workflow rather than an unordered list.

DimensionReasoningScore

Conciseness

The body is a lean seven-item list with no filler or explanation of concepts Claude already knows, matching the 'every token earns its place' anchor and clearly above the slightly-padded anchor at 2.

3 / 3

Actionability

Provides concrete parameter values (num_workers = min(8, os.cpu_count()), pin_memory=True, persistent_workers=True) and tool names, but never assembles them into an executable DataLoader snippet, leaving guidance incomplete relative to the copy-paste-ready anchor at 3.

2 / 3

Workflow Clarity

Items are listed clearly but as an unordered tip set rather than a sequenced process, with no validation or verification step, matching the 'steps listed but checkpoints missing' anchor and below the explicit-checkpoint anchor at 3.

2 / 3

Progressive Disclosure

A short (under 50 lines) single-purpose skill with one heading and a well-organized numbered list and no external references needed, which the rubric allows to score 3 on progressive disclosure.

3 / 3

Total

10

/

12

Passed

Description

75%Weight 40%Scale 1-3

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 is well-structured with both a clear capability statement and an explicit 'Use when' trigger, giving strong completeness and distinctiveness. It is slightly held back by listing only one action and by a somewhat narrow trigger-term vocabulary.

Suggestions

Expand specificity by listing multiple concrete actions (e.g., tune worker count, pin memory, profile I/O) instead of the single phrase 'Optimize data loading pipeline'.

Broaden trigger terms to include natural variations users might say, such as 'dataset', 'data pipeline', 'GPU starvation', or 'slow training'.

DimensionReasoningScore

Specificity

Names the domain and one concrete action ("Optimize data loading pipeline to prevent GPU starvation") but does not enumerate multiple distinct actions, so it falls short of the multi-action anchor at 3 and above the vague single-phrase anchor at 1.

2 / 3

Completeness

Explicitly states what it does ("Optimize data loading pipeline to prevent GPU starvation") and when to use it ("Use when setting up DataLoader or data preprocessing"), satisfying both halves with an explicit trigger clause.

3 / 3

Trigger Term Quality

Includes natural terms a user would say ("DataLoader", "data preprocessing") but omits common variations like "dataset", "data pipeline", or "GPU starvation", matching the 'some relevant keywords but missing variations' anchor.

2 / 3

Distinctiveness Conflict Risk

Targets a clear niche (PyTorch DataLoader/data-loading performance) with specific triggers unlikely to fire for unrelated skills, matching the 'clear niche with distinct triggers' anchor.

3 / 3

Total

10

/

12

Passed

Validation

100%

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

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
aiming-lab/AutoResearchClaw
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.