fix: skip readability after explicit article match

[code-yeongyu]

Summary

Explicit article containers are now returned before falling back to Mozilla Readability parsing. This preserves the Tistory-style article body that the direct selector already identified, even if Readability would fail or choose surrounding page chrome.

Changes

• Return the explicit article candidate before invoking Readability.parse().
• Add a regression test that mocks Readability to throw and verifies explicit article conversion still succeeds.

QA / Evidence

• npm run check: typecheck and Biome passed.
  • Artifact: /Users/yeongyu/local-workspaces/senpi/.omo/evidence/webfetch-tistory-worktree-pr/external-npm-run-check-after-explicit.txt
• npm test: full Vitest suite passed, 4 files / 35 tests.
  • Artifact: /Users/yeongyu/local-workspaces/senpi/.omo/evidence/webfetch-tistory-worktree-pr/external-vitest-full-after-explicit.txt
• Focused regression: npm test -- test/webfetch-explicit-article.test.ts passed.
  • Artifact: /Users/yeongyu/local-workspaces/senpi/.omo/evidence/webfetch-tistory-worktree-pr/external-explicit-article-green.txt

Risks

Low. The change only short-circuits after an existing explicit article selector has already passed the minimum text threshold; generic pages still use the existing Readability fallback.

Secret safety

Evidence contains command output only. No tokens, auth headers, cookies, or credential-bearing logs are included.

---

Summary by cubic

Return explicit article containers before running @mozilla/readability to preserve Tistory-style article bodies and avoid failures when Readability mis-parses or throws. Fixes cases where page chrome was selected or parsing failed.

• Bug Fixes
  • Short-circuit to the explicit article when it meets the text threshold; skip Readability.parse().
  • Add a regression test that mocks @mozilla/readability and verifies skipping while producing the expected markdown.

^{Written for commit a5a65a7. Summary will update on new commits.}