CtrlK
BlogDocsLog inGet started
Tessl Logo

advanced-evaluation

Advanced Evaluation skill from sickn33/agentic-awesome-skills

44

Quality

44%

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 ./plugins/AI-Agents-Safe-Coding-Skills/skills/advanced-evaluation/SKILL.md

The canonical home for this skill is advanced-evaluation in administrakt0r/AI-Agents-Safe-Coding-Skills

SKILL.md
Quality
Evals
Security

Quality

Content

63%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 thorough, well-organized guide to LLM-as-a-judge evaluation with strong actionable templates and examples, but it is longer than necessary and references internal documents that do not exist as bundle files. Tightening the prose and either providing or removing the phantom internal references would raise its quality.

Suggestions

Move the three worked examples and the API-style JSON schemas into separate reference files under references/ (e.g. examples.md, prompt-templates.md) and link to them from a leaner SKILL.md overview.

Either create the three named 'Internal reference' files or remove those entries so signaled references resolve to real artifacts.

Trim the ASCII pipeline diagram, the redundant 'Guidelines' recap, and the boilerplate Metadata/Integration/Limitations sections to reduce token overhead.

DimensionReasoningScore

Conciseness

The body is information-dense and mostly efficient (taxonomy, bias table, metric matrix, prompt templates), but carries padding that could be trimmed: a large ASCII pipeline diagram, a redundant 'Guidelines' section restating points already made, and boilerplate 'Skill Metadata'/'Limitations'/'Integration' sections. This matches 'Mostly efficient but includes some unnecessary explanation or could be tightened'; it is below 4 because the padding is more than minor.

3 / 5

Actionability

It provides copy-paste-ready prompt templates for direct scoring and pairwise comparison, concrete JSON output schemas, a position-swap protocol, and three worked examples with full input/output JSON. This fits 'Mostly executable guidance; concrete code or commands with minor gaps'; it stops short of 5 because the prompt bodies contain placeholders ({prompt}, {criteria}) and the 'References' internal items are not linked to usable artifacts.

4 / 5

Workflow Clarity

Multi-step processes are clearly sequenced: the position-bias mitigation protocol has explicit numbered steps with a consistency-check feedback loop, and the pipeline diagram lays out ordered stages. This matches 'Clear sequence with most checkpoints present; minor validation gaps'; it is not 5 because the rubric-generation and scaling workflows lack explicit error-recovery checkpoints, and there is no validation step for the panel/hierarchical scaling flows.

4 / 5

Progressive Disclosure

The skill is well-sectioned but monolithic (~450 lines in a single SKILL.md) with no bundle files present, while the 'References' section names three 'Internal reference' items (LLM-as-Judge Implementation Patterns, Bias Mitigation Techniques, Metric Selection Guide) that are not real files — signaled references that resolve to nothing. This fits 'Some structure but could be better organized; references present but not clearly signaled; content that should be separate is inline'; the under-50-line simple-skill exception does not apply given the length.

3 / 5

Total

14

/

20

Passed

Description

25%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 is a terse source attribution rather than a capability statement: it names the domain and provenance but conveys no concrete actions, no natural trigger terms, and no usage guidance. It functions more as a label than as a trigger description.

Suggestions

Replace the attribution-style description with a capability statement, e.g. 'Designs LLM-as-a-judge evaluation pipelines: builds rubrics, mitigates position/length bias, and scores model outputs with confidence calibration.'

Add an explicit 'when' clause: 'Use when building automated eval pipelines, comparing model responses, or creating scoring rubrics for LLM outputs.'

Include natural trigger terms users would say (e.g. 'evaluating model outputs', 'LLM judge', 'rubric', 'pairwise comparison') instead of the generic word 'Evaluation'.

DimensionReasoningScore

Specificity

The phrase 'Advanced Evaluation skill' names the domain (evaluation) but offers no concrete actions — it describes what the skill is rather than what it does (e.g. 'designs evaluation pipelines', 'scores model outputs'). This matches the anchor 'Names the domain but actions are minimal or generic', and falls well below score 3 which requires 1-2 concrete actions.

2 / 5

Completeness

It gives only a vague 'what' ('Advanced Evaluation skill') and has no 'when' / 'Use when...' clause at all; per the judging guidelines a missing 'Use when...' clause caps completeness at 3, and the near-total absence of both what and when places it at 2 ('Has a vague what and no when'). It does not reach 3 because there is no clear statement of concrete capability.

2 / 5

Trigger Term Quality

The only keyword is 'Evaluation', which is generic and technical; it lacks natural phrases a user would actually say ('evaluating model outputs', 'LLM-as-a-judge', 'rubric'). This fits 'One or two generic keywords; missing the natural phrases users say' and is below score 3 which expects some relevant keywords with synonyms.

2 / 5

Distinctiveness Conflict Risk

'Advanced Evaluation skill' is very broad and would overlap with any other evaluation-related skill (foundational 'evaluation', metric skills, etc.), matching 'Very broad; high overlap risk with many similar skills'. It is above 1 only because it does name a specific domain rather than being wholly generic like 'Helps with code and documents'.

2 / 5

Total

8

/

20

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
administrakt0r/AI-Agents-Safe-Coding-Skills
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.