agent-evaluation
Testing and benchmarking LLM agents including behavioral testing, capability assessment, reliability metrics, and production monitoring—where even top agents achieve less than 50% on real-world benchmarks
33
30%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/agent-evaluation/SKILL.mdQuality
Discovery
32%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 identifies a clear domain (LLM agent testing and benchmarking) and lists relevant subcategories, but it reads more like a topic overview than an actionable skill description. It lacks a 'Use when...' clause, concrete actions the skill performs, and the statistical claim about benchmark performance is editorial rather than functional. The description needs explicit trigger guidance and more specific action verbs.
Suggestions
Add an explicit 'Use when...' clause with trigger scenarios, e.g., 'Use when the user asks to evaluate, benchmark, or test an LLM agent, or needs help designing evals, test suites, or reliability metrics for AI systems.'
Replace category labels with concrete actions, e.g., 'Designs behavioral test suites, creates capability benchmarks, calculates reliability metrics, and sets up production monitoring for LLM agents.'
Remove the editorial claim about '50% on real-world benchmarks' as it doesn't help Claude decide when to use the skill, and instead add natural trigger terms like 'eval', 'agent evaluation', 'test harness', 'LLM performance'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (LLM agent testing/benchmarking) and lists several areas (behavioral testing, capability assessment, reliability metrics, production monitoring), but these are more like categories than concrete actions. It doesn't specify what the skill actually does—e.g., 'generates test suites', 'runs benchmarks', 'produces reliability reports'. | 2 / 3 |
Completeness | The description addresses 'what' at a high level but completely lacks a 'Use when...' clause or any explicit trigger guidance for when Claude should select this skill. Per the rubric, a missing 'Use when...' clause caps completeness at 2, and since the 'what' is also somewhat vague, this scores a 1. | 1 / 3 |
Trigger Term Quality | Includes relevant terms like 'LLM agents', 'benchmarking', 'testing', 'reliability metrics', and 'production monitoring', which are reasonably natural. However, it misses common user variations like 'evaluate', 'eval', 'agent evaluation', 'test harness', 'accuracy', or 'performance testing'. | 2 / 3 |
Distinctiveness Conflict Risk | The focus on LLM agent testing and benchmarking is a fairly specific niche, but the broad terms like 'testing', 'monitoring', and 'capability assessment' could overlap with general software testing skills or monitoring/observability skills. The statistical claim about 50% benchmarks adds flavor but doesn't help disambiguation. | 2 / 3 |
Total | 7 / 12 Passed |
Implementation
27%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is extremely verbose, embedding massive illustrative code blocks inline that explain patterns Claude could infer from much shorter descriptions. The code is not truly executable (missing types, unimplemented methods) reducing actionability, and the entire content is monolithic with no progressive disclosure. The conceptual coverage is broad but the token cost is very high relative to the actionable value delivered.
Suggestions
Reduce each pattern to 20-30 lines max: state the key insight, show one minimal executable example, and link to a separate file for the full implementation.
Make code examples truly executable by choosing a real framework (e.g., PromptFoo, Langsmith SDK) and providing copy-paste ready snippets with actual imports and runnable commands.
Split the four major patterns and four sharp edges into separate bundle files (e.g., patterns/statistical-evaluation.md, sharp-edges/benchmark-production-gap.md) and keep SKILL.md as a concise overview with links.
Add explicit validation checkpoints to workflows, e.g., 'After establishing baseline, verify statistical significance with at least N=30 runs before proceeding to regression testing.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~600+ lines. Massive code blocks explain concepts Claude already knows (statistical testing, chi-squared tests, Jaccard similarity). The interfaces and classes are illustrative pseudocode that could be condensed to patterns and key decisions. Sections like 'What is a PDF' equivalent explanations of basic testing concepts waste tokens. | 1 / 3 |
Actionability | The code examples are TypeScript-like but not truly executable—they reference undefined types (Agent, AgentOutput, AgentContext, TestCase), missing imports, and unimplemented helper methods (containsRudeLanguage, isRelevantToCustomerService, etc.). They illustrate patterns but aren't copy-paste ready. No concrete tool commands or real framework integration shown. | 2 / 3 |
Workflow Clarity | The Collaboration section has brief workflow sequences (design → create suite → implement → evaluate → iterate), but these are high-level and lack validation checkpoints. The patterns themselves show some flow (run tests → analyze → identify concerns) but there are no explicit verification steps or error recovery loops for the overall evaluation process. | 2 / 3 |
Progressive Disclosure | Monolithic wall of text with no bundle files or external references. All content is inline—hundreds of lines of code that should be in separate reference files. No links to detailed documentation, examples files, or API references. The skill would benefit enormously from splitting patterns into separate files with a concise overview in SKILL.md. | 1 / 3 |
Total | 6 / 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.
Validation — 9 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
skill_md_line_count | SKILL.md is long (1136 lines); consider splitting into references/ and linking | Warning |
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 9 / 11 Passed | |
- Repository
- sickn33/antigravity-awesome-skills
- Commit
8854d4e
- 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.