[dmccoystephenson]

Summary

Stage B (unit-test expansion) cycle targeting check_placeholders in scripts/check_docs.py — the check that enforces CLAUDE.md's "every {{placeholder}} has a Step 4 row" rule. Six fixture cases are added to CheckPlaceholdersTests, each characterizing current behavior; no production code is changed.

• Undeclared {{#if FLAG}} errors — the conditional-flag failure branch had only a happy-path case. The error names {{FLAG}} and never reports the closing {{/if}}.
• Rows after the Step 4 section do not count — a matching row placed under the next ### heading still leaves the placeholder reported missing.
• Step 4 as the last section — with no following ### heading, the table is read to end of file.
• Multiple missing placeholders — reported in one error, sorted ({{ALPHA}}, {{ZED}}).
• Whitespace inside braces — {{ NAME }}, {{#if FLAG }}, and {{ /if }} are trimmed before matching.
• Unused table rows are not reported — only the template → table direction is enforced. This is documented as current behavior rather than treated as a bug: the check_docs.py docstring and CLAUDE.md state only the forward direction. A probe of the live file found no genuinely unused row today; the only row with no template use is the path/to/file example row in the DOC_SOURCES_TABLE guidance (line 651), so a naive reverse check would false-positive on it.

No tracking issue — gap found during triage (Phase 2 Stage B: the failure branches of check_placeholders had no fixture case).

Research grounding

RESEARCH.md §3 — regression guards on behavior-bearing code. No template or phase definition is changed, so no Implementations entry applies.

Doc sync check

No docs, template, Step, or placeholder changed. python3 scripts/check_docs.py passes.

Test plan

• python3 -m unittest discover -s tests -v — Ran 19 tests, OK (13 before this change)
• python3 scripts/check_docs.py — "Doc consistency check passed."
• /create-dev-loop behavioral run — not applicable (test-only change; nothing the generated skill receives is touched)

Skipped issues

This cycle was spent on test expansion rather than the open backlog, for the same reasons recorded in #153:

• CLAUDE.md still links gardener under the old dmccoystephenson owner #106, RESEARCH.md Implementations entries record a ship date that is unknowable when the entry is written #128 — need an edit to CLAUDE.md, which is harness-blocked in this loop and needs separate maintainer authorization.
• [self-audit] The repo's own validation anchor is structurally unrunnable in a headless dispatch, so template changes can never merge autonomously #109, Unverified: whether an in-checkout .worktrees/ survives gardener's refresh (git clean -fdx) #114, Phase 4 verifies backward 'Closes #N' references but not forward 'a follow-up issue has been filed' promises #133, Phase 5's read-only 'gh api .../pulls/<n>/comments --jq' is refused by the attribution hook as an unsigned post #135, Upstream COMMANDS.md is trusted as source, but it can disagree with the command class it documents #141, A blanket merge pre-authorization does not resolve a do-not-auto-merge path match, and Phase 8 does not say so #142, Stage A sweeps fix the instances named, not the class — and a prose grep is treated as a CI-strength anchor #143, grep is not on tend's allow-list, so generated validation checklists and verify-by-removal assertions cannot run headlessly #127 — observations about how the loop or gardener behaves headlessly. Their fixes belong in the skill repo or gardener, or need a human-run anchor.
• Phase 4's rubric scores the diff only, so PR-description obligations like CLAUDE.md's retrofit note go unchecked #120–Negated closing keywords in PR bodies still auto-close issues #125, An approval-gated CI run reads as 'no checks reported', which is neither the green nor the red Phases 1 and 4 branch on #129–Test the mechanism, not the stochastic outcome, when a mock cannot exercise an acceptance criterion #131, Tests-fix has no path for a scope-blind anchor, only for an anchor that cannot run #140, Phase 1 triage never checks whether the default branch's own CI is green #144–Phase 5: compress the review wait for a single autonomous run with no reviewer assigned or requested #150 (template-rule) — each adds a new rule to the template body. Nothing automated tests the template body, and a headless session cannot do the real /create-dev-loop validation run. Phase 5: compress the review wait for a single autonomous run with no reviewer assigned or requested #150 is partly covered by Name the Copilot reviewer @copilot; skip the review poll when no reviewer was requested #154's "Reviewer addition fails" change, but its "no reviewer configured" case is still open.

This PR description was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

🤖 Generated with Claude Code

---

drafted by Claude on behalf of Daniel Stephenson

[dmccoystephenson]

Self-review rubric (scored against CI run 37901826523 and the diff):

• Scope: PASS — the diff touches only tests/test_check_docs.py, +42 lines, all inside CheckPlaceholdersTests.
• Tests-new: PASS (n/a) — no new production function; six new cases target the existing check_placeholders.
• Tests-fix: n/a — no bug fix. Each new case was checked by reasoning against the mutation it should catch. If #if prefix stripping were dropped, the error would name {{#if FLAG}}, which does not contain the asserted {{FLAG}}. If /if were not skipped, the assertNotIn("/if", ...) would fire. If the table section were widened to the whole file, LATER would be declared and the expected error would be missing.
• Sibling structure: PASS — the new cases reuse the class's _table helper and setUp reset, and assert on message content per the module docstring.
• Sibling renames: n/a — nothing renamed.
• Docs: PASS — no documented behavior changed. The unused-row case records a limitation that the check_docs.py docstring already implies (it only enforces the forward direction).
• Issue resolution: n/a — Stage B cycle, no Closes.
• CI: PASS — doc-consistency is green on the PR head. Locally: 19 tests ran, OK (13 before this change), and the doc consistency check passed.
• Placeholder parity / README Step parity / Fence escaping / Phase numbering / Version comments / No back-ported specifics / gardener interface / Retrofit note: n/a — create-dev-loop.md and README.md are untouched.
• Research grounding: PASS — the PR body cites RESEARCH.md §3.
• Behavioral validation: UNVERIFIED-not-applicable — the change is test-only.

Judgment flag: test_unused_table_row_is_not_reported pins the current one-directional behavior. A future reverse-direction check would need to update this case, and would also need to exempt the path/to/file example row at create-dev-loop.md:651. The test comment marks this as characterization, so a deliberate change is not blocked.

Summary: no blocking findings. Merge-ready once the review poll completes.

This comment was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

---

drafted by Claude on behalf of Daniel Stephenson

[Copilot]

🟢 Approval recommended

The focused tests accurately characterize the existing implementation and introduce no unresolved issues.

0 open findings

What changed in this PR

Adds regression coverage for check_placeholders without changing production behavior.

Changes:

• Adds six fixture cases covering conditionals, section boundaries, sorting, whitespace, and unused rows.
• Confirms existing one-way template-to-table validation behavior.

File	Description
tests/​test_check_docs.py	Expands check_placeholders unit-test coverage.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.