scan_packages: baseline sentencepiece dup2 finding after upstream reindent

[danielhanchen]

What

Add the current-indentation variant of the sentencepiece os.dup2 finding to the scan_packages.py allowlist so the supply-chain scan stops blocking on it.

Why it was failing

pip scan-packages :: hf-stack and pip scan-packages :: studio both fail on the "Scan declared + transitive Python deps" step with a single non-baselined CRITICAL:

[1] CRITICAL  Reverse shell / bind shell pattern
    Package:  sentencepiece
    File:     sentencepiece/__init__.py
    Evidence: L772: os.dup2(self.ostream.fileno(), self.orig_stream_fileno)
            | L777: os.dup2(self.orig_stream_dup, self.orig_stream_fileno)

This is a benign false positive: sentencepiece redirects stdout/stderr file descriptors to capture its C++ logs, and the heuristic flags os.dup2 because reverse shells dup socket fds onto 0/1/2. No socket or networking is involved.

The finding is already allowlisted, but the baseline key is (package, package-relative file, check, evidence_hash), where evidence_hash is over the matched code with the L<NN>: markers stripped and the code's own indentation preserved. A newer sentencepiece release reindented this block, moving it from L1221/L1226 to L772/L777 and changing its leading indentation. That changed the hash (bba233.. to 65b5a11c..), so the existing entry no longer matched and the finding resurfaced as a blocking CRITICAL. The scan gate only trips on non-baselined CRITICAL/HIGH, which is why the MEDIUM findings are unaffected.

Fix

Add the new indentation variant to scripts/scan_packages_baseline.json. The old L1221 entry is kept so both sentencepiece versions stay covered.

Verified with the scanner's own _load_baseline / _finding_key: the L772/L777 finding key is now present in the baseline set, and the old L1221 variant is still covered.

[chatgpt-codex-connector]

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.
Repo admins can enable using credits for code reviews in their settings.

[gemini-code-assist]

Code Review

This pull request adds a baseline exception for the 'sentencepiece' package in 'scripts/scan_packages_baseline.json' to ignore a critical severity 'Reverse shell / bind shell pattern' check. There are no review comments, so I have no feedback to provide.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.