AI coding agents

Coding agents write better apitest tests when they know the library's conventions and its traps. Two things help: the apitest skill, and these docs in Markdown.

The apitest skillCopy Anchor Link

The apitest repository includes an Agent Skill that teaches an agent how to write, fix and review apitest tests. It covers:

• The shape of a test, and the rules that prevent the most common mistakes, such as a missing End() or a mock without a Status.
• How mocks match requests, and why a client with its own Transport bypasses them.
• The JSONPath traps, such as numbers decoding as float64.
• Debugging failed tests, table-driven tests, shared configuration and reports.

The full API is in a reference file that the agent reads only when it needs it, so the skill stays small until it is used.

Claude CodeCopy Anchor Link

Add the apitest repository as a plugin marketplace, then install the plugin:

claude plugin marketplace add steinfletcher/apitest

claude plugin install apitest@apitest

Inside a session, /plugin marketplace add steinfletcher/apitest followed by /plugin install apitest@apitest does the same.

Claude uses the skill on its own whenever a task involves apitest, and /apitest:apitest runs it directly. To fetch the latest version of the skill, run:

claude plugin update apitest@apitest

Other agentsCopy Anchor Link

The skill follows the open Agent Skills format, so agents that support it can use the folder as it is. Copy skills/apitest from the repository into the agent's skills directory. For a single project in Claude Code, that is .claude/skills/apitest:

git clone --depth 1 https://github.com/steinfletcher/apitest.git /tmp/apitest

cp -r /tmp/apitest/skills/apitest .claude/skills/apitest

These docs as MarkdownCopy Anchor Link

Every page of this site is also available as Markdown, with the code samples included:

Each page also has a Copy Markdown button below its title, and an Open menu that sends the page to ChatGPT, Claude, Cursor or Scira AI.