fix: Stringify gen_ai.tool.output attribute

[alexander-alderman-webb]

Description

Issues

Reminders

• Please add tests to validate your changes, and lint your code using tox -e linters.
• Add GH Issue ID & Linear ID (if applicable)
• PR title should use conventional commit style (feat:, fix:, ref:, meta:)
• For external contributors: CONTRIBUTING.md, Sentry SDK development docs, Discord community

[github-actions]

Semver Impact of This PR

🟢 Patch (bug fixes)

📋 Changelog Preview

This is how your changes will appear in the changelog.
Entries from this PR are highlighted with a left border (blockquote style).

---

New Features ✨

• (ci) Cancel in-progress PR workflows on new commit push by joshuarli in #5994
• Add db.driver.name spans to database integrations by ericapisani in #6082

Bug Fixes 🐛

• (google_genai) Redact binary data in inline_data and fix multi-part message extraction by ericapisani in #5977
• (grpc) Add isolation_scope to async server interceptor by robinvd in #5940
• (profiler) Stop nulling buffer on teardown by ericapisani in #6075

• Stringify gen_ai.tool.output attribute by alexander-alderman-webb in #6086

Internal Changes 🔧

• (celery) Remove unused NoOpMgr from utils by sentrivana in #6078
• (pydantic-ai) Remove dead Model.request patch by alexander-alderman-webb in #5956
• (tests) Replace deprecated enable_tracingwith traces_sample_rate by sentrivana in #6077
• Set explicit base-branch for codecov action by ericapisani in #5992

---

_{🤖 This preview updates automatically when you update the PR.}

[github-actions]

Codecov Results 📊

✅ 27 passed | Total: 27 | Pass Rate: 100% | Execution Time: 4.51s

📊 Comparison with Base Branch

Metric	Change
Total Tests	—
Passed Tests	📈 +2
Failed Tests	—
Skipped Tests	📉 -2

✨ No test changes detected

All tests are passing successfully.

❌ Patch coverage is 40.00%. Project has 15422 uncovered lines.
❌ Project coverage is 27.82%. Comparing base (base) to head (head).

Files with missing lines (3)

File	Patch %	Lines
langchain.py	3.25%	⚠️ 595 Missing
utils.py	14.71%	⚠️ 87 Missing
execute_tool.py	0.00%	⚠️ 25 Missing

Coverage diff

@@            Coverage Diff             @@
##          main       #PR       +/-##
==========================================
- Coverage    28.41%    27.82%    -0.59%
==========================================
  Files          190       190         —
  Lines        21354     21366       +12
  Branches      7068      7068         —
==========================================
+ Hits          6067      5944      -123
- Misses       15287     15422      +135
- Partials       558       557        -1

---

Generated by Codecov Action

[sentry-warden]

safe_repr with set_data_normalized double-quotes string tool outputs

Using safe_repr(output) when output is a string adds Python repr quotes (e.g., "hello" becomes "'hello'"). The existing test at line 736 of test_langchain.py expects to call int() on the tool output - this will fail if the output "5" becomes "'5'". Other integrations use safe_repr directly with span.set_data(), not with set_data_normalized() which already handles type conversion.

Verification

Traced through safe_repr (sentry_sdk/utils.py line 562-566) which calls repr() on values. For strings, repr() adds quotes. Verified set_data_normalized (sentry_sdk/ai/utils.py lines 492-499) passes strings through unchanged. Confirmed existing test at tests/integrations/langchain/test_langchain.py:736 calls int() on tool output, which would fail with quoted strings.

Identified by Warden code-review · RLV-WJ8