skill-testing

A testing workflow for Agent Skills using cases written in YAML, a plain-text format for structured data. It checks whether a skill triggers correctly, follows its output rules, and handles unusual inputs.

In plain words
What is it for?
Use it to run a skill's test cases, evaluate expected versus actual behavior, or produce a skill test report.
Why use it?
It provides repeatable dry-run tests and captures the agent's responses, making it easier to find behavior or formatting problems in a skill.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/wangwei1237/agent-skills/skill-testing
Any agent
npx skills add wangwei1237/agent-skills --skill skill-testing
Clone the repo
git clone --depth 1 https://github.com/wangwei1237/agent-skills

Made for: Claude Code, Codex.

Per session 61 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,670 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00061 $0.02670
Opus 5 $0.00030 $0.01335
Sonnet 5 $0.00012 $0.00534
Haiku 4.5 $0.00006 $0.00267

Measured 2d ago against content hash c7d63aea20ab, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-testing scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/skill_test_runner.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skill-testing/SKILL.md · 202 lines

How it starts

The opening of the file, as written. The whole thing — 202 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill Testing

Use this skill to evaluate another Skill from its tests/skill_cases.yaml file. The test target is Skill behavior, not real business execution. Keep every case in dry-run mode unless the YAML explicitly says otherwise.

Default behavior is end-to-end. A user request to test a Skill or run a tests/skill_cases.yaml file authorizes isolated sub-agent capture for the test cases: generate dry-run prompts, run each rewritten prompt in a fresh isolated sub-agent, capture the actual dry-run response, format it, score it, and return the final Skill Test Report in the same turn.

Inputs

Accept either:

  • A Skill directory containing SKILL.md and tests/skill_cases.yaml.
  • A direct path to a skill_cases.yaml file.
  • A YAML case file plus Agent responses captured from a separate harness.

The default response directory is the target Skill's tests/tmp/response/ directory. For example, skills/my-skill/tests/skill_cases.yaml uses skills/my-skill/tests/tmp/response/.

If the user provides a response directory explicitly, consume those response files instead of generating new responses. Otherwise, create or replace response files in the default response directory with sub-agent outputs and continue to final scoring.

Workflow

  1. Load the YAML test file.
  2. Validate required top-level fields: version, skill.name, defaults.mode, defaults.execution, and cases.
  3. Validate every case has id, type, title, query, inputs.files, inputs.context, expected.skill, expected.must_include, expected.must_not_include, expected.artifacts, expected.routing, and assertions.
  4. Build one dry-run prompt per case from the original query.
  5. Unless an explicit response directory was provided, dispatch each rewritten prompt to a fresh sub-agent in dry-run mode and capture the sub-agent's final response text. The sub-agent must receive only the rewritten prompt plus the dry-run restriction, not the YAML expected fields, assertions, scoring rubric, or report from other cases.
  6. Save the raw response under <target-skill>/tests/tmp/raw_response/<case_id>.md.
  7. Format the raw response into the Skill Test response contract and save one scorer-facing response file per case in <target-skill>/tests/tmp/response/<case_id>.md.
  8. Run the scorer against the formatted response directory.
  9. Output a deterministic final report with case-level status and an overall result.

Read the full file on GitHub · 202 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 202 lines · 61 tokens per session scan A c7d63aea20ab

Subscribe to this mod's changes

skill-testing is a skill published in the GitHub repository wangwei1237/agent-skills (4 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 61 tokens to every session and 2,670 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

nature-statistics

Audit, revise, or draft manuscript statistical reporting for Nature / high-impact journal submissions. Use when the user asks to check statistical analysis sections, p values, confidence intervals, sample size, biological versus technical replicates, randomization, blinding, multiple-comparison correction, model…

Yuan1z0825/nature-skills · 139 tokens

nature-citation

Add strict Nature/CNS citations to manuscript text by splitting long passages into citable segments, searching only accepted flagship and subjournal titles from Nature Portfolio, the AAAS Science family, and Cell Press, filtering by publication time range, and exporting one reference-manager-ready output by default.…

Yuan1z0825/nature-skills · 236 tokens

baoyu-comic

Knowledge comic creator supporting multiple art styles and tones. Creates original educational comics with detailed panel layouts and batch-capable image generation. Use when user asks to create "知识漫画", "教育漫画", "biography comic", "tutorial comic", or "Logicomix-style comic".

JimLiu/baoyu-skills · 61 tokens

baoyu-youtube-transcript

Downloads YouTube video transcripts/subtitles and cover images by URL or video ID. Supports multiple languages, translation, chapters, and speaker identification. Caches raw data for fast re-formatting. Use when user asks to "get YouTube transcript", "download subtitles", "get captions", "YouTube字幕", "YouTube封面"…

JimLiu/baoyu-skills · 107 tokens

cesiumjs-camera

CesiumJS camera control - Camera, flyTo, lookAt, setView, ScreenSpaceCameraController, CameraEventAggregator, flight animation. Use when positioning the camera, creating flyTo animations, constraining user navigation, tracking entities, or converting between screen and world coordinates.

CesiumGS/cesiumjs-skills · 61 tokens

cesiumjs-custom-shader

CustomShader authoring — vertexShaderText and fragmentShaderText against VertexInput, FragmentInput, FeatureIds, Metadata, czmmodelMaterial. Use when reading EXTmeshfeatures or EXTstructuralmetadata property textures/tables, vertex displacement, or shading VoxelPrimitive.

CesiumGS/cesiumjs-skills · 61 tokens