test: align candidate capability count

Author: parousia8888

docs/V0.5.4_ENGINEERING_PLAN.md

@@ -40,8 +40,8 @@ opens a lower-friction support channel and submits the one currently eligible cu
 | G1 | Rule-contract conformance rename | Runtime, tests, generated JSON/Markdown, npm files and capability surfaces use the new name; no active public surface presents planted fixtures as ground truth | completed |
 | G2 | Real-world regression corpus | One deterministic command executes structured cases for report rendering, pnpm workspace inheritance, nested templates, moved-condition retest and benign numeric SVG review; generated JSON/Markdown records provenance and limits | completed |
 | G3 | First-trial and v0.5.4 candidate surfaces | Main-branch first trial is unpinned; CI/release/trusted install examples stay pinned; candidate version surfaces agree on 0.5.4 while published release state remains 0.5.3 | completed |
-| G4 | Focused and final verification | New focused tests pass, generated artifacts are byte-current, one final `npm run check` and Skill Creator validation pass | in progress |
-| G5 | Authorized external actions | Discussions is enabled and verified; awesome-devsecops rules are rechecked and a factual v0.5.3 PR is submitted if still eligible | pending |
+| G4 | Focused and final verification | New focused tests pass, generated artifacts are byte-current, one bounded complete-suite pass and Skill Creator validation pass | completed |
+| G5 | Authorized external actions | Discussions is enabled and verified; awesome-devsecops rules are rechecked and a factual v0.5.3 PR is submitted if still eligible | in progress |
 | G6 | Release approval checkpoint | Candidate commits and external evidence are pushed; exact release operations and residual risks are presented to the author | pending |
 | G7 | Public v0.5.4 release | Signed tag/assets, GitHub Release, npm provenance, `v1` consumer verification and release-state update | blocked on explicit approval |
 
@@ -200,6 +200,18 @@ Run in this order:
 Do not add unrelated browser, network, live-target or multi-runtime matrices to this patch. CI and
 CodeQL remain the independent platform gates after push.
 
+Completion record (2026-08-16):
+
+- Skill Creator validation returned `Skill is valid!`.
+- The final check completed lint and tests through JS/TS audit, then exposed one stale test-output
+  literal: p7 expected 24 capabilities while the generated contract correctly contained 25 after
+  adding the non-detection regression capability.
+- Updated only that literal. To respect the verification budget, continued from p7 through every
+  not-yet-run test and ran shell smoke once instead of repeating the already-passed first half.
+- Result: all lint gates, all 44 test files and shell smoke passed across the bounded continuation;
+  the real-adapter test retained its documented opt-in skip. Remote CI will provide the independent
+  single-command `npm run check` replay after push.
+
 ## G5 - External actions
 
 Discussions:

test/p7-surfaces.test.mjs

@@ -9,7 +9,7 @@ const ROOT = fileURLToPath(new URL('..', import.meta.url));
 
 for (const [script, expected] of [
   ['generate-launch-evidence.mjs', /9 stable detection, 0 planned detection, 5 journeys, 5 studies/],
-  ['check-p7-surfaces.mjs', /tutorials, agent lifecycle, 24 capabilities, 5 journeys, 5 studies/],
+  ['check-p7-surfaces.mjs', /tutorials, agent lifecycle, 25 capabilities, 5 journeys, 5 studies/],
 ]) {
   const args = [join(ROOT, 'scripts', script)];
   if (script.startsWith('generate-')) args.push('--check');