State in design cards that every failing_tests method must be red at base #1002

New issue

Open

opened 2026-10-12 14:39:23 +00:00 by jared · 0 comments

jared commented

2026-10-12 14:39:23 +00:00

Owner

Problem

Takedown final-red requires every new test method in failing_tests to be red at base (#995, commit 7907c2b6). The design cards still say "failing_tests: only the tests that are red at base" and "Prefer a new test file", and treat each entry as a whole file. The cards do not say that every test method in a failing_tests file must be red at base. The cards do not say where pass-case and guard tests go.

Observed

Takedown run docs/takedown/1001-* (ticket #1001) failed at final-red. The designer put pass-case tests in the new failing test file. The critic and the grader did not object across 2 rounds. The run cost $0.71.

final design revise-1: failing test not red at base (not red at base: ResearchCauseTest#test_a_filled_class_and_a_mechanism_with_a_path_line_pass, ResearchCauseTest#test_a_path_line_inside_the_mechanism_passes)

Reproduce

n/a: the failure shows in the run record at docs/takedown/1001-*.

Approach

Each card states that every test method in a failing_tests file must be red at base. Pass-case and guard tests go in a separate test file under code_files.

Required work:

Add the rule to the blind-designer card — check: present every test method in plugins/os-sdlc/eval/ticket-takedown/harness/cards/blind-designer.md
Add the rule to the blind-designer-apply card — check: present every test method in plugins/os-sdlc/eval/ticket-takedown/harness/cards/blind-designer-apply.md
Add the rule to the reviser card — check: present every test method in plugins/os-sdlc/eval/ticket-takedown/harness/cards/reviser.md

Orchestrator work: This is a harness change. Per the eval rule it lands after Jared's feedback, so the ticket is not marked ready-for-agent.

Origin

Trigger: takedown run on #1001 (docs/takedown/1001-*)
Improvised this session: none
Chain: final-red failure on pass-case tests ← cards say "only the tests that are red at base" per file ← #995 (7907c2b6) tightened final-red to test-method level and did not update the cards. Stop: DESIGN (#995).
Root candidate: none
Where: plugins/os-sdlc/eval/ticket-takedown/harness/cards/blind-designer.md:47, blind-designer-apply.md:47, reviser.md:36
Session: df535dc7-379a-4f68-9c40-2f8430b79acb
Transcript: /home/jared/.claude/projects/-home-jared-dev-cc-os--claude-worktrees-prism-architect/df535dc7-379a-4f68-9c40-2f8430b79acb.jsonl

Related: #995, #1001.

VERDICT: CREATE

## Problem Takedown `final-red` requires every new test method in `failing_tests` to be red at base (#995, commit 7907c2b6). The design cards still say "`failing_tests`: only the tests that are red at base" and "Prefer a new test file", and treat each entry as a whole file. The cards do not say that every test method in a `failing_tests` file must be red at base. The cards do not say where pass-case and guard tests go. ## Observed Takedown run `docs/takedown/1001-*` (ticket #1001) failed at final-red. The designer put pass-case tests in the new failing test file. The critic and the grader did not object across 2 rounds. The run cost $0.71. ``` final design revise-1: failing test not red at base (not red at base: ResearchCauseTest#test_a_filled_class_and_a_mechanism_with_a_path_line_pass, ResearchCauseTest#test_a_path_line_inside_the_mechanism_passes) ``` ## Reproduce n/a: the failure shows in the run record at `docs/takedown/1001-*`. ## Approach Each card states that every test method in a `failing_tests` file must be red at base. Pass-case and guard tests go in a separate test file under `code_files`. Required work: 1. Add the rule to the blind-designer card — check: present `every test method` in `plugins/os-sdlc/eval/ticket-takedown/harness/cards/blind-designer.md` 2. Add the rule to the blind-designer-apply card — check: present `every test method` in `plugins/os-sdlc/eval/ticket-takedown/harness/cards/blind-designer-apply.md` 3. Add the rule to the reviser card — check: present `every test method` in `plugins/os-sdlc/eval/ticket-takedown/harness/cards/reviser.md` Orchestrator work: This is a harness change. Per the eval rule it lands after Jared's feedback, so the ticket is not marked ready-for-agent. ## Origin - Trigger: takedown run on #1001 (`docs/takedown/1001-*`) - Improvised this session: none - Chain: final-red failure on pass-case tests ← cards say "only the tests that are red at base" per file ← #995 (7907c2b6) tightened final-red to test-method level and did not update the cards. Stop: DESIGN (#995). - Root candidate: none - Where: plugins/os-sdlc/eval/ticket-takedown/harness/cards/blind-designer.md:47, blind-designer-apply.md:47, reviser.md:36 - Session: df535dc7-379a-4f68-9c40-2f8430b79acb - Transcript: /home/jared/.claude/projects/-home-jared-dev-cc-os--claude-worktrees-prism-architect/df535dc7-379a-4f68-9c40-2f8430b79acb.jsonl Related: #995, #1001. VERDICT: CREATE

jared added the

plugin/os-sdlc

filed-by/agent

category/prose-artifact

labels

2026-10-12 14:39:23 +00:00

jared referenced this issue

2026-10-12 14:45:57 +00:00

Set cause_challenged in the takedown run's final record #999

jared referenced this issue

2026-11-11 07:04:41 +00:00

Accept edited existing test methods in takedown final-red #1687