Add if: always() so Module 19 whoami-runner.yml prints its 'where did this run?' receipt even on a red build #41

New Issue

2026-06-22T14:23:57-04:00

claude commented

2026-06-22 14:23:57 -04:00

Problem

Track A's whole payoff is reading the runner's self-identification receipt, but that step runs after Lint and Test with no if: always(), so any ruff/pytest failure aborts the job before the receipt prints. By Module 19 tasks-app has accumulated changes from M15/17/18, so a non-green lint/test is plausible, and when it happens Track A produces no receipt at all.

Evidence

modules/19-runners-the-compute-behind-automation/lab/whoami-runner.yml: steps run ruff check . (~line 41) then pytest -q (~line 44) then - name: Where did this run? (~line 47) with no if condition. README Track A: "It runs the same lint-and-test …, then prints the runner's hostname."

Why it matters

Track A's mandatory deliverable (and its Check-for-understanding criterion) silently vanishes on any non-green run; gating a diagnostic behind unrelated checks is a real design defect.

Proposed change

Add if: always() to the "Where did this run?" step only (keep lint/test running — the file deliberately wants the runner to do real work). Mirror the README's existing pattern with a note on GitLab's when: always. Do not split jobs or remove the checks.

Acceptance criteria

The identity/receipt step prints even when lint/test fail; the job still reports red.
Lint and Test still run.

Affected files

modules/19-runners-the-compute-behind-automation/lab/whoami-runner.yml, modules/19-.../README.md (optional GitLab note)

References

Source finding F32 (realVotes 3/3).

Filed from an adversarial multi-agent course review (217 raw findings → 54 adversarially-verified survivors). Scoped for manual review; intentionally not auto-assigned to an agent.

## Problem Track A's whole payoff is reading the runner's self-identification receipt, but that step runs after Lint and Test with no `if: always()`, so any ruff/pytest failure aborts the job before the receipt prints. By Module 19 `tasks-app` has accumulated changes from M15/17/18, so a non-green lint/test is plausible, and when it happens Track A produces no receipt at all. ## Evidence `modules/19-runners-the-compute-behind-automation/lab/whoami-runner.yml`: steps run `ruff check .` (~line 41) then `pytest -q` (~line 44) then `- name: Where did this run?` (~line 47) with no `if` condition. README Track A: "It runs the same lint-and-test …, then prints the runner's hostname." ## Why it matters Track A's mandatory deliverable (and its Check-for-understanding criterion) silently vanishes on any non-green run; gating a diagnostic behind unrelated checks is a real design defect. ## Proposed change Add `if: always()` to the "Where did this run?" step only (keep lint/test running — the file deliberately wants the runner to do real work). Mirror the README's existing pattern with a note on GitLab's `when: always`. Do not split jobs or remove the checks. ## Acceptance criteria - [ ] The identity/receipt step prints even when lint/test fail; the job still reports red. - [ ] Lint and Test still run. ## Affected files - `modules/19-runners-the-compute-behind-automation/lab/whoami-runner.yml`, `modules/19-.../README.md` (optional GitLab note) ## References Source finding F32 (realVotes 3/3). --- *Filed from an adversarial multi-agent course review (217 raw findings → 54 adversarially-verified survivors). Scoped for manual review; intentionally not auto-assigned to an agent.*