Always Fail Kit

Purpose

A test instrument for the AHP platform. It fails on every run, so a solution's failure branch can be exercised end to end without contriving a broken kit.

Use it as the step an if/else node branches on: the step always fails, so the failure branch always runs and the success branch is always skipped.

What it does

Writes one line to its log explaining that the failure is deliberate, then exits with status 1. That is all. It reads nothing and produces nothing.

Inputs

None. The kit declares no file input pins and reads no environment variables of its own.

Output

The kit declares one output pin, result.txt, which it never writes. A step that fails produces no file, and that is the point: wire the pin to a downstream step and watch the consumer be skipped rather than fed.

There is no manifest file in this repository. The output pin is declared on the platform when the kit version is registered, alongside the image.

Logging

Written to /tmp/logs/always_fail_kit.log and to stdout.

Build and run locally

With uv

uv sync --extra dev
uv run python always_fail/execute.py; echo "exit status: $?"

With Docker

docker build -t always-fail-kit .
docker run always-fail-kit; echo "exit status: $?"

Either way the exit status is 1 and the log line lands in /tmp/logs/always_fail_kit.log.

Tests

uv run pytest