refactor(#334): rename /generative-engine-audit → /geo-audit#339
Conversation
- Directory move: .claude/skills/generative-engine-audit/ -> /geo-audit/ - AgDR-0043 renamed to -geo-audit-skill.md; H1 + chosen-option references updated; new Consequences entry documenting the post-v1 rename and the no-shim choice (industry GEO term + /seo-audit sibling shape). - Template renamed to templates/audits/geo-audit.md; H1 -> "GEO Audit" - Smoke test references updated (paths, name field, CLAUDE.md count 53, /geo-audit row assertions); all 51 checks pass. - CLAUDE.md skill table row rewritten (terse description, alphabetised by position, stays under 25-word Wave 1 invariant). - docs/multi-project.md row updated; trail-of-rename appended. - Cross-refs swept: launch-check (dispatch row + deep-dive prose), seo-audit (sibling see-also), ai-crawlers.json (description + see_also), site/skill.md, site/llms-full.txt, AgDR-0044 + AgDR-0045. - No backward-compat shim per ticket recommendation (Option A) — skill shipped < 1 week ago, no real adopter usage to preserve. Closes #334
atlas-apex
left a comment
There was a problem hiding this comment.
Code Review: PR #339
Commit: 2783f93dec435739c970136ece5d941b3fdf7bbe
Summary
Mechanical rename of the audit-family skill /generative-engine-audit → /geo-audit so the slash command matches the industry-standard GEO term and pairs cleanly with /seo-audit. Skill shipped < 1 week ago (PR #315), so Option A (clean rename, no shim) is the right call — no meaningful adopter usage to preserve. 13 files: directory + 2 file renames (git mv-tracked), targeted content updates, and a ### Post-v1 rename Consequences subsection added to AgDR-0043 capturing the date, ticket, no-shim choice, and rationale.
Checklist Results
- Architecture & Design: Pass — pure mechanical rename, no architectural change. AgDR-0043 shape preserved.
- Code Quality: Pass —
git mvused (similarity 80–93% on renamed files), targeted content edits, no incidental drift. - Testing: Pass — smoke test (
51/51) updated for new paths; Wave 1 invariants (test_token_efficiency_wave1.sh) all pass; both verified locally at this SHA. - Security: Pass — N/A (no auth / crypto / secrets / dependencies touched).
- Performance: Pass — N/A (no runtime code).
- PR Description & Glossary: Pass — four narrative bullets (each answering what + why), Glossary present with 9 well-defined terms,
Closes #334. - Technical Decisions (AgDR): Pass — AgDR-0043 already documents the design; rename decision recorded as a
### Post-v1 renameConsequences subsection (correct placement, not a new top-level section). PR description carries the Option A rationale explicitly. - Adopter Handbooks: N/A — public-layer always-load handbooks (
architecture/clean-architecture-layers.md,architecture/migration-safety.md,general/commit-message-quality.md) don't apply to a pure-rename, no-source-code, no-migration-path PR. Diff touches only.md,.sh,.json,.txt, so no language handbooks activate either. No custom-handbook layer configured for this fork.
Verification (run locally at HEAD)
| Claim | Verified |
|---|---|
Rename completeness — no stale /generative-engine-audit references in framework files outside deliberate trails |
YES — 6 remaining mentions are all in 3 documented trail locations: AgDR-0043 (4 mentions: > opener, axis-1 row, ### Post-v1 rename heading, body), AgDR-0044 line 9 (with parenthetical "renamed in #334"), docs/multi-project.md row 702 ("Originally shipped as … renamed in #334") |
| Old paths absent on disk | YES — .claude/skills/generative-engine-audit/, docs/agdr/AgDR-0043-generative-engine-audit-skill.md, templates/audits/generative-engine-audit.md all gone |
| AgDR-0043 Post-v1 rename note location | YES — H3 at line 119, under ## Consequences (line 95), same level as Positive / Negative / Migration. Not a new top-level section. |
| Wave 1 invariants | bash .claude/hooks/tests/test_token_efficiency_wave1.sh — all 4 pass. CLAUDE.md /geo-audit row description = 10 words (well under 25-word cap). SKILL.md description: = 127 chars (well under 200 hard cap). |
| Skill catalogued | YES — 53 skill dirs on disk; ### Available skills (53) in CLAUDE.md; /geo-audit row present. Count is correct (mutation-test added separately in PR #338 makes the 52→53 base). |
| Smoke test coverage | bash .claude/skills/geo-audit/tests/smoke.sh — 51/51 pass across 16 contract groups. No hard-coded old skill name in assertions; the 4 remaining generative-engine mentions are dimension-name checks (intentional — assertion 15 explicitly pins that the dimension label stays). |
| AI-crawler registry | JSON parses; 12 crawlers; see_also updated to geo-audit paths; description updated. |
| Landing-site touch | site/skill.md skill-list reflects new name; site/llms-full.txt skill description updated. |
| PR body quality | Narrative bullets per pr-quality.md, Glossary present, Closes #334 triggers auto-close on merge. |
| Ticket #334 | OPEN, properly named, labeled enhancement, P2. |
Issues Found
None.
Suggestions
None blocking. Two minor observations recorded for posterity:
/launch-checkretains "Generative-engine" as the dimension name (row 6 label) while the deep-dive skill is now/geo-audit. This is intentional and the smoke test pins it (assertion 15). Worth a follow-up someday — either rename the dimension to "GEO" to match the skill, or document the deliberate divergence in AgDR-0043. Not for this PR.projects/apexyard/audits/generative-engine-audit/persisted history (if any exists in private portfolios) keeps the old path by design — per AgDR-0019, persisted artefacts are immutable. The PR description correctly notes this is out of scope.
Verdict
APPROVED — clean, surgical rename with thorough cross-reference sweep, all tests pass, AgDR shape preserved, narrative-quality PR body. Ship it.
🤖 Reviewed by Rex (Code Reviewer Agent)
📌 Reviewed commit: 2783f93dec435739c970136ece5d941b3fdf7bbe
- Directory move: .claude/skills/generative-engine-audit/ -> /geo-audit/ - AgDR-0043 renamed to -geo-audit-skill.md; H1 + chosen-option references updated; new Consequences entry documenting the post-v1 rename and the no-shim choice (industry GEO term + /seo-audit sibling shape). - Template renamed to templates/audits/geo-audit.md; H1 -> "GEO Audit" - Smoke test references updated (paths, name field, CLAUDE.md count 53, /geo-audit row assertions); all 51 checks pass. - CLAUDE.md skill table row rewritten (terse description, alphabetised by position, stays under 25-word Wave 1 invariant). - docs/multi-project.md row updated; trail-of-rename appended. - Cross-refs swept: launch-check (dispatch row + deep-dive prose), seo-audit (sibling see-also), ai-crawlers.json (description + see_also), site/skill.md, site/llms-full.txt, AgDR-0044 + AgDR-0045. - No backward-compat shim per ticket recommendation (Option A) — skill shipped < 1 week ago, no real adopter usage to preserve. Closes #334 Co-authored-by: me2resh <ahmed.abdelaliem@gmail.com>
Summary
/generative-engine-audit→/geo-audit(introduced via PR feat(#311): /generative-engine-audit — LLM/agent SEO sibling to /seo-audit #315, AgDR-0043) so the slash command matches the industry-standard GEO (Generative Engine Optimization) term and the established sibling-shape with/seo-audit. Future operators searching for "GEO" land on the right skill; current operators get terser/launch-checkoutput tables (4-char skill name vs 22-char).### Post-v1 renameConsequences subsection records the date, ticket, no-shim decision, and the rationale (industry term + sibling shape + terseness).git mv; in-file references updated inCLAUDE.mdskill table,docs/multi-project.mdbehaviour table,/launch-check(dispatch row + 3 prose mentions),/seo-audit(sibling cross-link),.claude/registries/ai-crawlers.jsondescription + see_also,site/skill.md+site/llms-full.txt, and AgDR-0044 + AgDR-0045 historical references; smoke test rewritten with new paths and updated CLAUDE.md skill-count assertion (53 instead of 52).Files touched (13)
.claude/skills/generative-engine-audit/{SKILL.md,tests/smoke.sh}→.claude/skills/geo-audit/...docs/agdr/AgDR-0043-generative-engine-audit-skill.md→AgDR-0043-geo-audit-skill.mdtemplates/audits/generative-engine-audit.md→templates/audits/geo-audit.mdCLAUDE.mddocs/multi-project.md.claude/skills/launch-check/SKILL.md.claude/skills/seo-audit/SKILL.md.claude/registries/ai-crawlers.jsondocs/agdr/AgDR-0044-token-efficiency-wave-1.mddocs/agdr/AgDR-0045-mutation-test-skill.mdsite/skill.md,site/llms-full.txtGrep verification:
grep -r generative-engine-audit . --exclude-dir=.git --exclude-dir=projects --exclude-dir=.claude/worktreesreturns only the deliberate trail-of-rename references (Consequences section, "originally shipped as" footnotes in AgDR-0044 + multi-project.md row). Persisted audit artefacts underprojects/apexyard/audits/generative-engine-audit/are immutable historical records — out of scope per ticket.Testing
bash .claude/skills/geo-audit/tests/smoke.sh— 51 / 51 contract checks pass (frontmatter, all 6 check buckets, verbatim naming-clash callout, cross-links, AgDR shape + sections + GEO/AEO distinction, registry validity + 11 named crawlers, template H1 + Findings header, CLAUDE.md count + row, multi-project row, /seo-audit cross-link, /launch-check dispatch row + dimension description, 9-dimension sweep claim).bash .claude/hooks/tests/test_token_efficiency_wave1.sh— all 4 Wave 1 invariants pass (CLAUDE.md rows ≤ 25 words; SKILL.md description ≤ 200 chars hard / ≤ 120 chars soft; every skill catalogued in CLAUDE.md; SessionStart banners ≤ 600 chars).ls .claude/skills/ | grep -v ^_ | wc -l).generative-engine-auditin framework files.Closes #334
Glossary
llms.txt,AGENTS.md, JSON-LD), different consumer./geo-audit/seo-audit; fanned out from/launch-check./seo-audit/launch-check/geo-auditfor row 6 (Generative-engine)./generative-engine-audit). Records the sibling-vs-extend axis, theskill.mdvsSKILL.mdnaming-clash callout, and (as of this PR) the post-v1 rename trail..claude/skills/<name>/tests/smoke.sh— verifies documentation invariants (frontmatter, cross-links, table presence, etc.) that the skill depends on. Pins shape; catches drift.