fix(diagnostics): classify core error markers case-insensitively

[bingran-you]

Summary

• make classify_error() use the already-normalized lowercase error string for install failed and ACP error markers
• add regression coverage for capitalized install failures and lowercase ACP errors

Fixes #541.

Verification

• uv sync --extra dev --locked
• uv run python -m pytest tests/test_scoring.py -q
• uv run ruff check src/benchflow/_utils/scoring.py tests/test_scoring.py
• uv run ruff format --check src/benchflow/_utils/scoring.py tests/test_scoring.py
• uv run ty check src/benchflow/_utils/scoring.py

[bingran-you]

Automation validation (2026-07-02): focused fix for #541 is ready for review.

Validated on head 768f54e1:

• uv sync --extra dev --locked
• uv run python -m pytest tests/test_scoring.py -q (48 passed)
• uv run ruff check src/benchflow/_utils/scoring.py tests/test_scoring.py
• uv run ruff format --check src/benchflow/_utils/scoring.py tests/test_scoring.py
• uv run ty check src/benchflow/_utils/scoring.py

Merge gate: wait for GitHub CI plus normal non-author review before squash merge.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7a01e4e6a5

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Name the guarded commit in the regression docstring

The repo-level AGENTS.md requires regression tests to name the PR or commit they guard, but this new regression docstring only cites issue #541. Please include the guarding PR or commit identifier here so the test follows the documented convention and remains traceable to the exact fix.

Useful? React with 👍 / 👎.

[bingran-you]

Users Simulation follow-up (2026-07-02T16:45Z): current head 768f54e1 is ready for this simulation scope. Focused local checks passed (tests/test_scoring.py, ruff check/format, ty on the scoring utility), and GitHub checks are now green: test, pip-audit, manifest-parity, integration-scope detect-scope, and integration-light detect-scope + rollout-smoke. Labels remain status:ready + review:pending; normal non-author review gate still applies.