peekie

Peekie skill

peekie is a CLI for Xcode .xcresult bundles. Four data-axis subcommands, each emitting JSON (or list / sonar). Reach for it before composing xcrun xcresulttool calls — Peekie is faster, terser, and runs only the xcresulttool invocations its subcommand needs.

Subcommands

peekie tests     <xcresult>  [--format json|list|sonar]  default: list
peekie warnings  <xcresult>  [--format json|list]        default: json
peekie errors    <xcresult>  [--format json|list]        default: json
peekie coverage  <xcresult>  [--format json|list]        default: json

Each subcommand is single-purpose. To get both warnings and coverage, run two commands and pipe / merge their outputs.

When to use which

User goal	Subcommand
"Did the tests pass?" / "What failed?"	peekie tests
"What warnings are in this build?"	peekie warnings
"Why is the build broken?" / "Show me errors"	peekie errors
"What's the coverage?" / "Coverage per module"	peekie coverage
"Generate SonarQube test report"	peekie tests --format sonar --tests-path …

Output shapes

peekie tests --format json

{
  "coverage": null,
  "modules": [
    {
      "name": "Calculator",
      "suites": [
        {"name": "CalculatorTests", "tests": [{"name": "testAdd()", "status": "success", "durationMs": 12.5, "message": null}]}
      ],
      "files": []
    }
  ]
}

files and top-level coverage are empty/null because peekie tests doesn't fetch coverage data. Use peekie coverage for that.

peekie warnings (default JSON)

[
  {"file": "Foo.swift", "line": 42, "column": 8, "type": "DeprecatedDeclaration", "message": "'oldFoo()' is deprecated: use bar"},
  {"file": "Foo.swift", "line": 51, "column": 8, "type": "DeprecatedDeclaration", "message": "'oldFoo()' is deprecated: use bar"}
]

Flat sorted array. line / column are null when xcresult didn't emit them. Type values are the raw Apple issueType strings (Swift Compiler Warning, Swift Compiler Error, DeprecatedDeclaration, No-usage, plus anything new Apple adds).

peekie errors (default JSON)

Same shape as peekie warnings.

peekie coverage (default JSON)

{
  "coverage": 0.6234,
  "modules": [
    {
      "name": "Calculator",
      "coverage": 0.71,
      "coveredLines": 167,
      "totalLines": 233,
      "files": [
        {"name": "Calculator.swift", "path": "/…/Calculator.swift", "coverage": 0.71, "coveredLines": 167, "totalLines": 233}
      ]
    }
  ]
}

Recipes

CI gate: fail when there are build errors

errors=$(peekie errors Build.xcresult | jq 'length')
if [ "$errors" -gt 0 ]; then
  peekie errors Build.xcresult --format list
  exit 1
fi

Coverage badge value

peekie coverage Tests.xcresult | jq '.coverage * 100 | floor'

Group warnings by type

peekie warnings Build.xcresult | jq 'group_by(.type) | map({type: .[0].type, count: length})'

Send only failures to a chat

peekie tests Tests.xcresult --include failure,mixed --format json | …

Filtering

• peekie tests --include … — comma-separated statuses (success,failure,skipped,expectedFailure,mixed,unknown). Default: all.
• peekie warnings / peekie errors have no --include flag. Filter the JSON output with jq or grep the list output.

Installation

If peekie is not on $PATH, suggest the user run:

mise use github:dodobrands/peekie

Or download a binary from Releases. Do not suggest swift run peekie … — that's the contributor flow, not the user flow.

Don't

• Don't invent --include-coverage / --include-warnings / --include-tests flags on the CLI. They exist on the SDK; on the CLI the subcommand itself determines what gets fetched.
• Don't reach for xcrun xcresulttool directly when one of the four subcommands gives you the same data. peekie warnings runs one xcresulttool call; the equivalent hand-rolled pipeline runs three.
• Don't expect peekie tests to print coverage or peekie coverage to print suites — each subcommand is single-purpose by design.