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 Use when: agent testing, agent evaluation, benchmark agents, agent reliability, test agent.
68
61%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./.agent/skills/agent-evaluation/SKILL.mdQuality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong skill description that follows best practices. It clearly specifies what the skill does (testing/benchmarking LLM agents with four specific activities), includes an explicit 'Use when:' clause with natural trigger terms, and carves out a distinct niche. The description uses proper third-person voice and provides enough detail for Claude to select it appropriately.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'behavioral testing, capability assessment, reliability metrics, and production monitoring'. Also includes a concrete detail about benchmark performance ('less than 50% on real-world benchmarks'). | 3 / 3 |
Completeness | Clearly answers both what (testing and benchmarking LLM agents with specific activities listed) and when (explicit 'Use when:' clause with trigger terms). The structure follows the recommended pattern. | 3 / 3 |
Trigger Term Quality | Includes natural keywords users would say: 'agent testing', 'agent evaluation', 'benchmark agents', 'agent reliability', 'test agent'. These cover common variations of how users would phrase requests about testing LLM agents. | 3 / 3 |
Distinctiveness Conflict Risk | Clear niche focused specifically on 'LLM agents' testing and benchmarking, distinct from general testing skills or general LLM skills. The combination of 'agent' + 'testing/benchmarking' creates a specific domain unlikely to conflict with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
22%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill content is essentially a stub that delegates everything to sub-skills without providing any actionable guidance in the main file. It lacks concrete examples, executable code, workflow sequences, and meaningful overview content. The structure for progressive disclosure exists but the main skill fails to stand alone as useful documentation.
Suggestions
Add a 'Quick Start' section with at least one concrete, executable example of how to set up a basic agent evaluation (e.g., a simple behavioral test with code)
Include a workflow section that sequences when to use each sub-skill (e.g., '1. Start with behavioral contracts, 2. Add statistical evaluation, 3. Layer adversarial tests')
Remove or condense the narrative framing in the intro and replace with actionable guidance on what makes agent testing different (specific patterns, not philosophy)
Add brief descriptions next to each sub-skill link explaining what it covers and when to use it, so users can navigate without clicking through
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The introductory paragraphs contain some unnecessary narrative framing ('You're a quality engineer who has seen...') that doesn't add actionable value. The capabilities/requirements lists are efficient, but the intro could be tighter. | 2 / 3 |
Actionability | The skill provides no concrete code, commands, or executable guidance. It only lists sub-skill references and abstract descriptions without any actual implementation details, examples, or copy-paste ready content. | 1 / 3 |
Workflow Clarity | There is no workflow, sequence, or process defined. The content is just a list of links to sub-skills with no guidance on how to approach agent evaluation, what order to follow, or validation checkpoints. | 1 / 3 |
Progressive Disclosure | The skill does reference sub-skills appropriately (one level deep), but the main file provides almost no overview content—it's essentially just a table of contents with no quick-start guidance or context for when to use each sub-skill. | 2 / 3 |
Total | 6 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- Dokhacgiakhoa/antigravity-ide
- Commit
332e58b
- 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.