Skip to content

docs(tdd): cut the C listings from the anti-patterns and testable-architecture pages - #18035

Open
melodic-automation-lanes[bot] wants to merge 2 commits into
mainfrom
work-resolve-issue/issue-17767-44944165148-1
Open

melodic-automation-lanes[bot] wants to merge 2 commits into
mainfrom
work-resolve-issue/issue-17767-44944165148-1

Conversation

@melodic-automation-lanes

Copy link
Copy Markdown
Contributor

Closes #17767

Summary

Opened by the work-resolve-issue lane. The run's own summary:

Part 4 of 6 of #16758. Row count: 2. F0404 (R11, K7, anti-patterns-khorikov.md): CUT, all 11 C#/Moq CallBase/DateTime.Now fences removed; part of the page cut, not the whole page. F0405 (R11, K7, testable-architecture-khorikov.md): CUT, all 6 C#/xUnit.net/Fluent Assertions fences removed (the CRM before/after, the domain-layer [Fact] test, CanExecute/Execute, domain events and their assertion); part of the page cut, not the whole page. Neither row is GONE: every cited listing was still on main. Neither page is a routing site where tdd:principles takes a language-specific step, so the listings were deleted with no generic replacement code. Where a listing held the only statement of a teaching point, one language-neutral sentence keeps the page coherent (the code-pollution flag and its real/fake interface fix, partial mocking of a concrete class, the ambient time provider, the two ways to inject time, when domain events are recorded and sent, what CanChangeEmail returns). Two prose items that only served a removed listing were deleted with it (R11's consumer rule): the 'listing 11.17' sentence naming DateTimeServer and the Khorikov quote about using a string for errors. The raw object\[\] type in prose became 'raw database rows'. No whole-page deletion, so no whole-page predicted outcome applies; predicted outcome for the partial cuts: no loss, since every rule both pages teach is still stated in prose. Script readers: a Grep of the whole checkout (excluding .lane) for anti-patterns-khorikov|testable-architecture-khorikov found no script that reads either file by name: the hits are .performance/ratchets.json (the two ceilings, lowered here), docs/upstream/aihero-shipping-course.md:194 and docs/upstream/mattpocock-skills.md:49 (prose records), three CHANGELOG entries, plugins/tdd/skills/principles/SKILL.md:28,30 (routing table links), property-based-testing.md:31 (section-3 anchor, still present) and refactoring-under-test.md:87 (link). scripts/skill-stack-fences.py reads every skill .md by glob only to count fences. So both rows are cut and the Script readers list is empty. Eval case for tdd:principles that reads the cut text: none does. The closest existing case, plugins/tdd/evals/pr-interface-in-process (routes to tdd:principles; graders routes-principles, verdict-no, uses-real-calculator), asks about an in-process interface and may load anti-patterns-khorikov.md for code pollution, but none of its graders matches cut listing text; it is the candidate for the host session's targeted before-and-after case. No R4 or R8 row is in this part. AC10: no skill added.

Not verified

  • Tidy not run: the change spans 2 areas (plugins/tdd with .changes/tdd, and .performance), over the 1-area tidy cutoff.
  • check-affected skipped typos (not installed), comment-hygiene (fetch denied), size-check (script missing) and three validate-plugins sub-checks that run checked-tree code; left to the PR's CI.
  • python3 scripts/skill-stack-fences.py and the performance ratchet not executed here (no interpreter allowed in this lane); fence counts read by Grep. Left to the PR's CI.
  • JSON validity of .performance/ratchets.json checked by reading, not by a parser; left to the PR's CI.
  • Host-session eval (targeted before-and-after case on sonnet, do-not-merge hold and eval-pair record) is the host session's work per the issue, not this lane's.

Fix

  • plugins/tdd/skills/principles/reference/anti-patterns-khorikov.md: removed all 11 csharp fences (F0404); dangling prose reduced to language-neutral sentences; the 'listing 11.17' DateTimeServer sentence deleted.
  • plugins/tdd/skills/principles/reference/testable-architecture-khorikov.md: removed all 6 csharp fences (F0405); added one sentence stating what CanChangeEmail returns; deleted the string-for-errors quote that described the removed listing; 'object[]' and 'array length >= 3' reworded to raw…
  • .performance/ratchets.json: removed both files' skill-stack-fences ceilings (11 and 6), since both now count 0 and ratchet.py's tighten drops zero ceilings.
  • .changes/tdd/work-resolve-issue-issue-17767-44944165148-1-1162669f.md: new patch changelog fragment for tdd.

Verification

The runner ran each check from the base copy against 4eef3e9:

  • scripts/check-changelog-fragments.sh --check: passed
  • scripts/check-changed-skills.sh --merge-base: passed

The run reports:

  • Grep for fence openers in both pages: anti-patterns-khorikov.md has none; testable-architecture-khorikov.md keeps only the 4 untagged diagram fence lines (2 diagrams), which scripts/skill-stack-fences.py does not count, so both files count 0.
  • Grep for object[], Moq, xUnit, DateTime, CallBase, virtual in both pages: no match after the edits.
  • Grep for references to both file names across the checkout: no script reader; links and the section-3 anchor still resolve.
  • check-changelog-fragments: All 5 changelog fragments valid.
  • check-changed-skills: CHECK-SKILL principles PASS, 0 errors, 2 pre-existing warnings (no Gotchas surface, description 621 chars).
  • check-affected: 14 passed, 9 skipped over 4 changed files.
  • Fresh-context review subagent over the whole diff: 2 findings and 3 nits; fixed the dangling CanChangeEmail reference, the stale array-length precondition, the weak antecedent in anti-patterns section 1 and the changelog wording; left the dropped string-for-errors quote deleted, since it only…
  • The lane run: https://github.com/melodic-software/claude-code-plugins/actions/runs/44944165148

Left out: the items under Not verified.

Related

#17767

Lane-Activity: work-resolve-issue/resolve-issue
@github-actions

github-actions Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

claude-security-review reviewed c53544c47aba10eae5556093267f7aae7c3d6ab4: read 4 of 4 changed file(s); 0 finding(s), 0 as inline comments.

Security review: no security issues found

This review covered blocking issues only, because this lane already reviewed an earlier head (4eef3e9b). The new head is a merge of main into the branch. I read every hunk of all 4 changed files, and none of them matches a trusted path.

  • plugins/tdd/skills/principles/reference/anti-patterns-khorikov.md and testable-architecture-khorikov.md: the C# code listings are removed, and plain-language sentences replace them where a listing carried a point. These are skill reference pages, so I checked the deletions under the instruction-surface lens. I looked at what breaks once each removed piece of text is gone. Every removed block is a teaching example: no rule that guarded an action, granted a tool or blocked a request is removed. The practices the pages teach are still stated in prose, and none of the new text runs commands, grants tools or widens what the model may do. The protected-class register was not consulted; I judged these deletions by their consequences alone.
  • .performance/ratchets.json: removes two skill-stack-fences ceilings for files that now have no counted fences. This is a performance ratchet. It controls no trust, token, merge hold or kill switch.
  • .changes/tdd/…md: a new changelog fragment, prose only.

The diff touches no workflow, action, trust control, merge hold, plugin manifest or userConfig key. The PR context contains no instructions aimed at the lane. The only earlier post from this lane is a summary with no line anchor that reported 0 findings, so there is nothing to carry forward.

critical-findings: 0

@melodic-automation-lanes

melodic-automation-lanes Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor Author

Unattended merge: native auto-merge leaves this pull request unarmed at head 4eef3e9b2ec252e5ae0f656b554163448e1eb2db. No person needs to act: a lane clears this reason, and the next scheduled run judges it again.

pr-merge's gate holds it on its risk score (unreadable: the head is
  c53544c47aba10eae5556093267f7aae7c3d6ab4, not 4eef3e9b2ec252e5ae0f656b554163448e1eb2db)

@github-actions

github-actions Bot commented Nov 21, 2026 •

Copy link
Copy Markdown
Contributor

claude-review reviewed c53544c47aba10eae5556093267f7aae7c3d6ab4: read 4 of 4 changed file(s); 0 finding(s), 0 as inline comments.

Code review: no blocking findings. (Blocking-only review: this lane already reviewed an earlier head, 4eef3e9b.)

The new head c53544c4 merges main into the branch. The PR's own 4-file change is the same as before. It removes the 11 csharp fences from anti-patterns-khorikov.md and the 6 from testable-architecture-khorikov.md, rewrites the teaching points those listings carried as plain prose, drops both files' skill-stack-fences ceilings from .performance/ratchets.json, and adds a tdd patch changelog fragment.

What I checked at this head:

  • Diff. I read every hunk of all 4 changed files.
  • Remaining fences. A Grep at head finds no fences in anti-patterns-khorikov.md. testable-architecture-khorikov.md keeps only the two untagged diagram fences (lines 12–22 and 115–121), so dropping both ceilings fits a count of 0.
  • ratchets.json. The edited region still reads as well-formed JSON after the merge (lines 184–190). I checked it by reading, not with a parser.
  • Closing issue Cut the model-read ecosystem specifics from tdd: anti-patterns and testable-architecture #17767.
    • Both rows, F0404 and F0405, are cut, and the ceilings are removed rather than lowered, since 0 fences are left.
    • The other acceptance criteria are about the PR body: row count, script-reader search, eval case, whole-page versus partial cut, and AC10. The body addresses each of them.
    • The ratchet CI step is the check that would fail if tagged fences came back.
  • Author-claimed, not verified by me: the check-changelog-fragments, check-changed-skills and check-affected results and the fence counts stated in the PR body.

Not checked: the wrapper supplied no loosened_check_verdict:, loosened_check_patterns: or loosened_check_decision_record: lines. I used verdict fail with no consumer entries and did not check the citation requirement. The change loosens no check: the ceilings go only because the counts reached 0. Loosened checks would be marked: n/a.

No plugin manifest changed and no paved-path finding applies. The PR body and issue text contain no embedded instructions to the lane.

Verifier dropped 0 of 0 candidates.

@github-actions

Copy link
Copy Markdown
Contributor

Lane run report

Field Value
Activity pr-unblock/resolve-conflicts

Run 44950023717, attempt 1

Field Value
Model rung 0
Leg handoff
Run job success
Activity step success
Commits pushed none (read lane)
Run 44950023717

Failed GitHub API calls

No failed call recorded by the run job's scripts.

Lane-Activity: pr-unblock/resolve-conflicts
@github-actions

Copy link
Copy Markdown
Contributor

Lane run report

Field Value
Activity pr-unblock/apply-conflict-patch

Run 44950023717, attempt 1

Field Value
Model rung 0
Leg base
Run job success
Activity step success
Commits pushed c53544c
Run 44950023717

Failed GitHub API calls

No failed call recorded by the run job's scripts.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Cut the model-read ecosystem specifics from tdd: anti-patterns and testable-architecture

0 participants