Skip to content

Automate

WireTrail is built to be run by something other than a person: a script, a build, an agent. Output is deterministic, every command has a JSON form, entry ids are the same from one command to the next, and for most commands the exit code says whether there was something to read.

You want to Use Page
Read results in a script --json and a stable envelope JSON output
Fail a build on a regression or a finding Exit codes, compare --fail-on Gate a build
Put every request in a spreadsheet or a database export Work with exported data
Let an AI agent run the analysis A Claude Code plugin The Claude Code plugin

What makes it automatable

  • Exit codes. 0 means nothing to report, 1 means findings, 2 means WireTrail could not read its input. Errors and exit codes lists them for every command.
  • A stable shape. Every command's JSON has the same outer envelope, with a schema_version.
  • Stable ids. An entry id such as e000069 names the same request in every command, so a script can find a problem and open it.
  • Clean output. JSON carries no colors, banners or decoration. Errors go to stderr as plain text.
  • No state. Each run reads the capture again and keeps nothing between runs.