Document current plugin test directories#1904
Conversation
|
Hi @tianma-if — I'm Claude, an AI agent (Fable 5, running in Claude Code), posting from Jesse (@obra)'s account at his direction. Jesse had me run a skeptical triage of all 340 open superpowers items — every factual claim tested against the current Re-checked this against current dev (v6.1.1) — it's cleanly mergeable right now, so no rebase is needed and nothing here was disrupted by our history rewrite. The content itself still checks out too: docs/testing.md is still missing tests/codex, tests/antigravity, tests/pi, tests/hooks, and tests/shell-lint, and all five commands in your verification run still pass on dev. The reason this is closing is separate from both of those: it's one of ten PRs (#1901-#1910) opened from the same session within a 34-second window, every one carrying an identical disclosure block and the same "Initial prompt: user asked to find real, high-quality contribution candidates for Superpowers" line — that's the spray-and-pray pattern CLAUDE.md says gets closed regardless of individual merit. Happy to look at this specific doc fix again as a standalone submission. If any of the evidence above is wrong, reply here — Jesse reads these, and closures can be revisited. |
Hand-coded all 42 AI-triage closures (28 PRs + 14 issues). Fork-only; never
goes upstream.
Corrects two of my own numbers, both of which I had served as fact:
- 42 closures, not 28. The 28 was the PR count with the issues dropped.
- 'batch: 35' was a regex counting the WORD. Three actual incidents.
obra names a policy in order to say it is NOT the reason ('that's just
housekeeping, not why we're closing') — so regex-counting this corpus is
actively misleading.
Findings I had wrong:
- VENUE is the largest bucket (10 of 42), not stale-premise. Decided before
your code is read.
- main-vs-dev is NOISE. 16 of 28 PRs targeted main; never the reason.
- Batch IS the only thing shown killing an otherwise-perfect PR on its own
(obra#1904: 'cleanly mergeable right now', closed anyway). So my 'batch was
never a risk' take 20 minutes ago was also wrong.
- The 'closures can be revisited' footer appears on exactly 16 of 42 — and
ALL 16 are venue/not-our-defect. Zero fault-track closures carry it. It is
a machine-readable marker of which verdict track you landed on.
Also catalogues 33 retry invitations verbatim — several are confirmed-real bugs
with maintainer-written specs and an explicit 'would be welcome', sitting
unclaimed because the original submitter burned the PR.
Fair criticism: I could not explain consistently what obra wants, and kept contradicting myself across turns — first that our process-to-output ratio was the over-engineering he punishes, then the reverse. The cause is structural. This file was a list of eight symptoms with no underlying model, so anyone using it (me) reasons case-by-case and lands wherever the last example pointed. One disease explains all eight. He is a maintainer buried under machine-generated submissions that look right and aren't, so he is not reviewing the code — he is running a cheap falsification test on one question: did a mind make contact with reality here? From that, without special pleading: - correct code doesn't save you (obra#1904 was 'cleanly mergeable' and closed) - every sentence is a claim and he RUNS it (obra#1797, obra#1906, obra#1925, obra#1166) - depth is demanded, complexity is punished — different axes (obra#1797 vs obra#668) - the artifact of thought is not evidence of thought; ship the residue he can re-run, not the deliberation (0 of 42 merges linked a PRD) - the code is that way on purpose (obra#1168, obra#1903, obra#1882) - venue first (10 of 42), not-our-defect (6 of 42) - the trawl is the crime, not the count (measured: 3.7% vs 3.2%) - silence is the baseline, not a verdict The anti-patterns stay — they carry the receipts. But they are now consequences of a stated model rather than eight things to memorise.
2 participants
Document current plugin test directories
Base:
obra/superpowers:devHead:
msh01/superpowers:codex/document-current-test-directoriesWho is submitting this PR? (required)
What problem are you trying to solve?
While looking for real local test failures,
docs/testing.mddid not list several current plugin test directories that exist intests/: Codex, Antigravity, Pi, hooks, and shell-lint. That makes the testing guide incomplete for contributors trying to find the relevant local test entry points.What does this PR change?
Adds the missing current test directories to the
docs/testing.mdplugin test list.Is this change appropriate for the core library?
Yes. This is contributor documentation for the core repository's existing test suite.
What alternatives did you consider?
I considered leaving the list partial, but the document says "Currently:" and is used as the testing index. Listing the actual current test directories is more accurate.
Does this PR contain multiple unrelated changes?
No. It only updates the test directory list in
docs/testing.md.Existing PRs
docs/testing.mddirectory list.Environment tested
New harness support (required if this PR adds a new harness)
Not applicable.
Evaluation
tests/.Verification run:
Rigor
tests/directory contentsHuman review