> ## Documentation Index
> Fetch the complete documentation index at: https://docs.peepsai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Find tests and check coverage

> Search your Peeps test cases, read one in full, browse folders and script versions, and check what the code in front of you already has covered.

Everything on this page only reads. Nothing changes in Peeps until you approve a change on another page.

## Search your test cases

> Do we have a test for guest checkout?

Your agent searches test case titles and steps in your project. Each match comes back with its title, its full steps, the folder it is filed in, and a link to it in Peeps.

* **Search is literal.** It matches short text in titles and steps, ignoring case. Your agent tries several short terms and alternate wording rather than one long phrase.
* **No match is not proof of a gap.** An empty search only means those words were not found. Your agent says when coverage is unknown rather than claiming it is missing.
* **A test with no script is flagged.** A test case that exists but has no published script is coverage on paper only. No run can reach it.
* **Archived tests are hidden by default.** Say "including archived" to find one to restore.

You can also narrow a search to a folder, or to tests that are in no folder:

> List the tests in the Payments folder.

## Read a test case in full

> Show me the steps for the password reset test.

Your agent reads the complete test case: its details, all of its steps, the tests that run before it and after it, and its cleanup test. Nothing is summarized or cut short.

Secret values from your environments are masked wherever they appear. An archived test case is still readable in full.

## Browse folders and scripts

> What folders does this project have?

You get every folder, with how many test cases and subfolders each holds, and how many test cases are in no folder.

> Which script version does the login test run?

Your agent lists the script versions of a test case, newest first, and marks the one a run would use. It can then read that version's source, with secret values masked.

## Check coverage for your change

> Do we have Peeps tests for what I changed on this branch?

This is the `peeps-cover` workflow. Your agent reads the behaviors out of your local code and searches Peeps for each one, using several phrasings and paging through every result. It sends **search terms only** — your source never leaves your machine.

You get one proposal with each behavior marked:

| Mark          | Meaning                                                                                                                        |
| ------------- | ------------------------------------------------------------------------------------------------------------------------------ |
| **Unchanged** | An existing test still describes it correctly after your change.                                                               |
| **Stale**     | A test covers it, but its steps look out of date. Your agent describes what would need to change. It doesn't apply it.         |
| **Gap**       | The searches and the tests your agent read support that nothing covers it.                                                     |
| **Uncertain** | The evidence is incomplete. Your agent says what is missing and asks you — for your team's term for it, or for a test case id. |

Your agent prefers **uncertain** to a confident wrong answer, because a wrong "unchanged" is how a regression ships. It only proposes new tests for gaps.

Read the proposal rather than assuming it is complete, because search is literal. Nothing is created until you approve it. To act on the gaps, see [create and import tests](/create-and-import).

## Worked example: is this covered?

> **You:** I've added a "resend code" link to two-factor sign-in. Is that covered in Peeps?

Your agent reads the change in your checkout and searches Peeps for "two-factor", "2FA", "verification code" and "resend". It reports:

* **Unchanged:** "Sign in with a verification code" still covers the main flow.
* **Gap:** nothing covers resending a code.

It proposes one new test case for the gap, with draft steps, and asks whether to create it.