chore(deps): bump beautifulsoup4 from 4.14.3 to 4.15.0

[dependabot]

Bumps beautifulsoup4 from 4.14.3 to 4.15.0.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

[amit-gazal-thenvoi]

✅ Approved by the dependabot-triage skill.

Validated in a consolidated batch run: this bump was applied together with the
other passing Dependabot bumps (#492 opentelemetry-resourcedetector-gcp, #493
pillow, #494 fastapi, #497 click, #498 beautifulsoup4) on one local branch via a
single uv lock, so it's proven compatible with the rest of the batch — not just
green on its own. The lock diff contains exactly those 5 packages, with no
collateral drift, and uv lock --check passes.

Gate results with all 5 in place:

Check	Result
pre-commit run --all-files (ruff, ruff format, pyrefly, secrets, actionlint)	✅ passed
Unit suite	✅ 4083 passed, 82 skipped
crewai in an isolated dev-crewai venv (CI's exact target list) + pyrefly check	✅ 80 passed, 5 skipped, 0 errors
markdown-docs (CI's git ls-files '*.md' set)	✅ 57 passed
Integration (live platform)	✅ no bump-attributable failures
e2e baseline, core lane	✅ no bump-attributable failures

Integration and e2e were run for real, and every failure was attributed rather
than assumed:

• Integration — ran the full suite twice, once with the batch and once with
  only these 5 packages downgraded. The same 5 tests fail both ways
  (test_claude_sdk_approval ×2; test_letta_live ×3, APIConnectionError with
  no local Letta server), so they're pre-existing. The differing failures swap
  between runs (test_google_adk_converter ×4 and a WS 429 one way,
  test_control_signals ×2 the other) — ordering/live-platform flakiness. The
  google_adk group passes on the bumped versions when run in isolation.
• e2e core lane — 43 passed, 0 failed up to test_parlant_replies, which
  fails because parlant's own server needs ~6 min to become ready and the adapter
  gives up at 300s. A/B'd explicitly: with fastapi 0.140.0 startup took 6m03s and
  failed; with fastapi 0.136.0 it took 6m03s and failed identically. Not the bump.
• Out-of-lane adapters (copilot_acp, letta, google, crewai, opencode)
  skip under BAND_E2E_LANE=core and are covered by their own CI lanes; locally
  they lack the server/CLI/key regardless.

Safe to merge.