agent-skills-eval

agent-skills-eval

A test runner for Agent Skills.

Write a SKILL.md, drop in some evals, and find out — empirically — whether your skill actually makes the model better at the task.

Documentation · Quickstart · SDK · agentskills.io

Why this exists

Agent Skills — the open standard from Anthropic for giving agents domain knowledge — make it easy to ship a SKILL.md and assume your agent is now better at the task. The hard part is proving it.

agent-skills-eval is the missing piece. It runs your skill against the same prompts twice — once with_skill loaded into context, once without_skill (baseline) — has a judge model grade both outputs, and gives you a side-by-side report. If the skill doesn’t make a measurable difference, you’ll see it. If it does, you have receipts.

It’s the test framework for the Agent Skills ecosystem, separated from any specific agent runtime so it works wherever your skills do.

Quickstart

npx agent-skills-eval ./skills \
  --target gpt-4o-mini \
  --judge gpt-4o-mini \
  --baseline \
  --strict

That’s it. Point it at a folder of skills, give it a target model and a judge model, and it produces a workspace with full artifacts and a static HTML report.

agent-skills-workspace/
└── iteration-1/
    ├── meta.json            # run metadata
    ├── benchmark.json       # rolled-up pass/fail per skill
    ├── eval-basic/
    │   ├── with_skill/      # output, timing, judge grading
    │   └── without_skill/   # ↑ same, with the skill stripped
    └── report/
        └── index.html       # the visual report

Open iteration-1/report/index.html and you have a real, evidence-backed answer to “is my skill working?”

What you get

Install

npm install agent-skills-eval

Or run directly without installing:

npx agent-skills-eval --help

How it works

The mental model is straightforward. For every eval defined in your skill:

                ┌─────────────────────────────┐
                │       same prompt           │
                └───────────────┬─────────────┘
                                │
                ┌───────────────┴─────────────┐
                ▼                             ▼
        ┌──────────────┐              ┌──────────────┐
        │ with_skill   │              │without_skill │
        │ SKILL.md in  │              │ baseline,    │
        │ context      │              │ no skill     │
        └──────┬───────┘              └──────┬───────┘
               │                             │
               ▼                             ▼
          target model                  target model
               │                             │
               ▼                             ▼
            output                        output
               │                             │
               └──────────┬──────────────────┘
                          ▼
                   ┌─────────────┐
                   │  judge      │  scores both against
                   │  model      │  the same assertions
                   └──────┬──────┘
                          ▼
                  pass / fail per side

The judge sees the eval’s expected_output and assertions and grades each side independently. The --baseline flag is what enables the comparison; without it you only get the with_skill run.

YAML config

For anything beyond a quick command, drop a config file at the root of your project:

# agent-skills-eval.yaml
root: ./skills
workspace: ./agent-skills-workspace
baseline: true
target: gpt-4o-mini
judge: gpt-4o-mini
baseUrl: https://api.openai.com/v1
apiKeyEnv: OPENAI_API_KEY
include:
  - "skills/**"
exclude:
  - "**/draft-*"
evalIds:
  - "basic"
concurrency: 4
layout: iteration
strict: true
report:
  enabled: true
  title: Agent Skills Report
logging:
  format: pretty   # pretty | jsonl | silent
  verbose: false
  color: auto
targetParams:
  temperature: 0
judgeParams:
  temperature: 0

OPENAI_API_KEY=... npx agent-skills-eval --config agent-skills-eval.yaml

CLI flags always override config values.

Options reference

These are configuration-file options and CLI defaults. See --help for the available CLI flags; targetParams, judgeParams, and logging.snippetLength are config-only. Logging flags use --log-format, --log-file, --verbose, and --no-color; report flags use --report, --no-report, --report-title, and --report-output. Supplied CLI flags override matching config values.

To iterate on one case, pass --eval-id one or more times:

OPENAI_API_KEY=... npx agent-skills-eval --config agent-skills-eval.yaml --eval-id basic

SDK

For programmatic use — CI pipelines, custom dashboards, multi-skill rollups — drive the evaluator from TypeScript:

import {
  OpenAICompatibleProvider,
  consoleReporter,
  evaluateSkills,
} from "agent-skills-eval";

const provider = new OpenAICompatibleProvider({
  baseUrl: "https://api.openai.com/v1",
  apiKey: process.env.OPENAI_API_KEY!,
  model: "gpt-4o-mini",
  providerName: "openai",
});

const result = await evaluateSkills({
  root: "./skills",
  workspace: "./agent-skills-workspace",
  baseline: true,
  concurrency: 4,
  workspaceLayout: "iteration",
  strict: true,
  target: { model: provider.model, provider },
  judge: { model: provider.model, provider },
  onEvent: consoleReporter(),
});

console.log(result);

Stream events to a file as JSONL for downstream analysis:

import { jsonlReporter } from "agent-skills-eval";

const reporter = jsonlReporter({ file: "./events.jsonl" });

await evaluateSkills({ /* ... */ onEvent: reporter.onEvent });
await reporter.close();

Load YAML config programmatically:

import { loadConfigFile } from "agent-skills-eval";

const config = loadConfigFile("./agent-skills-eval.yaml");

Custom providers

Bring any backend by implementing the Provider interface — five fields, one method:

import type { Provider, ProviderResult } from "agent-skills-eval";

export const provider: Provider = {
  name: "my-provider",
  model: "my-model",
  async complete(prompt: string): Promise<ProviderResult> {
    return {
      provider: "my-provider",
      model: "my-model",
      output: "model output",
      latencyMs: 0,
      inputTokens: 0,
      outputTokens: 0,
      costUsd: 0,
    };
  },
};

Useful for: local model servers (Ollama, vLLM, llama.cpp), proprietary internal APIs, mock providers in unit tests, or routing layers in front of multiple providers.

Skill layout

A skill is a folder. The minimum is a SKILL.md. Add evals/evals.json and you can evaluate it.

my-skill/
├── SKILL.md
├── references/
│   └── notes.md
├── scripts/
│   └── helper.sh
└── evals/
    ├── evals.json
    └── files/
        └── input.csv

SKILL.md:

---
name: my-skill
description: Analyze small CSV files.
license: MIT
compatibility: Works with text-capable chat models.
---

When given a CSV file, identify the most important trend and cite the
relevant rows.

evals/evals.json:

{
  "skill_name": "my-skill",
  "evals": [
    {
      "id": "basic",
      "name": "basic behavior",
      "prompt": "Use the attached data to summarize revenue.",
      "files": ["evals/files/input.csv"],
      "expected_output": "The response identifies the highest revenue month.",
      "assertions": [
        "The output identifies the highest revenue month."
      ]
    }
  ]
}

If you skip assertions but provide expected_output, the SDK promotes the expected output into a judge assertion automatically — so a minimal agentskills.io eval file produces meaningful pass/fail grading without extra work.

CLI options

npx agent-skills-eval [root] \
  --config agent-skills-eval.yaml \
  --workspace ./agent-skills-workspace \
  --baseline \
  --target gpt-4o-mini \
  --judge gpt-4o-mini \
  --base-url https://api.openai.com/v1 \
  --api-key-env OPENAI_API_KEY \
  --include "skills/**" \
  --exclude "**/draft-*" \
  --eval-id basic \
  --concurrency 4 \
  --layout iteration \
  --strict \
  --log-format pretty \
  --report

Logging modes: pretty for humans, jsonl for machines, silent for quiet CI.

Reports

The static HTML report is built from disk artifacts and shows everything you’d want for skill iteration:

• Pass rate by skill and by eval
• Assertion-by-assertion grading evidence with judge reasoning
• Full target output, side by side for with_skill and without_skill
• Prompt and judge prompt details
• Timing and token usage
• Tool calls when present

Use --report-output (or report.output in YAML) to choose where the report lands.

agentskills.io compatibility

Implements the agentskills.io specification end to end:

• SKILL.md YAML frontmatter — required name and description, optional license, compatibility, metadata, allowed-tools
• Strict validation: name length, lowercase-hyphenated format, parent-directory match, description length, compatibility length
• Optional scripts/, references/, and assets/ directories — markdown references included in skill context, scripts exposed by manifest
• evals/evals.json schema: skill_name, evals[].id, prompt, expected_output, files, assertions
• Official artifact layout: iteration-N/<eval>/<mode>/outputs, timing.json, grading.json, benchmark.json
• Baseline comparison via with_skill and without_skill

Beyond the spec, this SDK adds: per-eval defaults, model params, tool definitions, deterministic tool_assertions, and a flat workspaceLayout: "flat" for multi-skill dashboards.

Examples

See examples/basic-skill for a complete skill folder, and examples/agent-skills-eval.yaml for a reference config.

Development

npm ci
npm test
npm pack --dry-run

Documentation

Full docs live at darkrishabh.github.io/agent-skills-eval (sources in docs/). Local preview:

python3 -m http.server 8080 --directory docs

Contributing

Issues, PRs, and skill examples are all welcome. See CONTRIBUTING.md, CODE_OF_CONDUCT.md, and SECURITY.md.

License

MIT. See LICENSE.

Integration boundaries

See the artifact contract for output layouts, field semantics, adapter guidance, and provenance limits. External result-format adapters can consume these files without adding dependencies to this runner.

The built-in OpenAI-compatible provider makes a single chat-completion request per target run (plus retries). Function tool definitions let you grade the returned tool-call requests; the provider does not connect to MCP servers, execute tools, or run a multi-turn agent loop. To evaluate a skill that needs real MCP execution today, implement a custom Provider that owns the tool loop and returns actual output and captured toolCalls for grading. Declaring a function tool is not evidence that its operation executed successfully.

This runner evaluates behavior supplied by a provider; it does not authenticate agent identity, detect clones, or attest to an untampered runtime. Integrations that need identity verification must perform it before starting the evaluation and retain the evidence separately. Passing scores do not establish identity or regulatory compliance.

Outgoing requests from OpenAICompatibleProvider carry User-Agent: agent-skills-eval/<package-version> (+https://github.com/darkrishabh/agent-skills-eval; node/<node-version>). This applies to target and judge calls and retries. SDK callers can override it through extraHeaders using any capitalization of User-Agent.

The proposed native-agent lifecycle is available for contributor review through the experimental runtime foundation. It is a separate SDK entry point; the native adapters and CLI integration are still under development with the authors of PRs #25 and #35.