CtrlK
BlogDocsLog inGet started
Tessl Logo

evaluating-llms-harness

lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.).

56

Quality

64%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/mlops/evaluation/lm-evaluation-harness/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.

The content is highly actionable with executable commands and well-organized progressive disclosure to real reference files, but it loses points on conciseness (redundant intros) and workflow clarity (no integrated validation checkpoints despite clear sequencing).

Suggestions

Remove the redundant restatement of "evaluates LLMs across 60+ academic benchmarks" between the "What's inside" and "Quick start" sections.

Add explicit verification checkpoints inside workflows (e.g., after running an evaluation, confirm results JSON exists and metrics are non-zero before proceeding).

Tighten the per-workflow checklists so they add sequencing detail beyond restating the step headers.

DimensionReasoningScore

Conciseness

The body is mostly actionable code and commands, but redundant intros ("What's inside" and the Quick start lead both repeat "evaluates LLMs across 60+ academic benchmarks") and checklists that restate step headers add padding that could be tightened, matching the score-2 anchor.

2 / 3

Actionability

It provides abundant copy-paste-ready, fully executable `lm_eval` commands plus complete bash and Python scripts with concrete model/task arguments, matching the score-3 anchor.

3 / 3

Workflow Clarity

The four workflows are clearly sequenced with step checklists, but they lack explicit validation or verification checkpoints inside the workflow, and error recovery lives only in a separate "Common issues" section rather than as integrated feedback loops, matching the score-2 anchor.

2 / 3

Progressive Disclosure

The body is an overview that clearly signals four one-level-deep references (benchmark-guide.md, custom-tasks.md, api-evaluation.md, distributed-eval.md) under "Advanced topics", all of which exist as real files, matching the score-3 anchor.

3 / 3

Total

10

/

12

Passed

Description

57%

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 terse and reasonably distinct, naming the tool and recognizable benchmark terms, but it lacks a "Use when..." trigger clause and broader action/keyword coverage, which together keep most dimensions at the mid level.

Suggestions

Add an explicit trigger clause, e.g. "Use when benchmarking model quality, comparing models on standard tasks, or reporting academic metrics like MMLU/GSM8K."

Expand the action list beyond a single verb ("benchmark") to concrete capabilities such as "evaluate, compare, and track models across standard benchmarks."

Include common user phrasings like "evaluate" or "eval" alongside "benchmark" to broaden trigger-term coverage.

DimensionReasoningScore

Specificity

The description names the domain ("lm-eval-harness") and one action ("benchmark LLMs") plus example benchmarks ("MMLU, GSM8K"), but it does not list multiple distinct concrete actions, so it falls short of the score-3 anchor.

2 / 3

Completeness

It states what the skill does ("benchmark LLMs") but includes no "Use when..." clause or equivalent trigger guidance, which per the judging guidelines caps completeness at 2.

2 / 3

Trigger Term Quality

"benchmark LLMs" and "MMLU, GSM8K" are natural terms a user would say, but common variations like "evaluate", "eval", or "test model" are missing, matching the score-2 anchor of relevant keywords without full coverage.

2 / 3

Distinctiveness Conflict Risk

The tool name plus specific benchmark triggers ("MMLU, GSM8K") carve out a clear academic-benchmarking niche unlikely to conflict with other skills, matching the score-3 anchor.

3 / 3

Total

9

/

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.