[pierreraby]

Summary

Carries #165's shared-redactor fix (8a2dc73) by @patlux / Patrick Wozniak, credited as co-author. Runtime source changes match the original commit; recalibrated on main after the other eight carries and the separate deterministic-abort test (#196).

• Remove the narrower duplicate credential patterns in runtime diagnostics in favor of redactCommandCodeErrorText.
• Preserve the prior URL query/fragment/userinfo removal before applying the shared redactor.
• Add a preservation assertion for unknown query parameters and URL userinfo, in addition to the original synthetic secret regression.

This is defensive hardening, not a confirmed production leak. It does not import or depend on open diagnostics PR #180. That carry adds content-free diagnostics, not raw error text storage; retain both additions when integrating runtime/test changes.

Validation

• Original synthetic-secret regression fails against main (standalone sk-shaped token is not removed).
• Final head e7cd59e329df867f7ff19e6d713d052d60e503eb on main 590ef638c119539bc72718998d501a6872b608cf: full PI_LOCAL_REQUIRED=1 OMP_COMPAT_REQUIRED=1 npm test PASS, including the deterministic abort test and real Pi/OMP mock integration.
• Final source range-diff against the original carry is identical (only changelog context adapted); all other integrated source and both RPC sequences are preserved.
• Format and diff checks PASS.
• Final active LSP probe of runtime and its test: no errors; two auxiliary hints on unchanged runtime code (unknown boundary parameter / large class).
• The project typecheck excludes the full host entrypoint; the known 17 inherited expanded-probe diagnostics are not fixed here.

The original nine-carry independent review covered this unchanged source patch. The separate abort stabilization received a focused DeepSeek V4.1 Flash review and re-review (OK). No real secrets, live login, billed requests, packed-artifact smoke or release. #165 is currently closed non-merged; the maintainer has chosen to retain that status after integration and add the final carry summary there.

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None

[github-actions]

Memory benchmark

Compared base 590ef63 with PR head e7cd59e on the same GitHub-hosted darwin runner. Lower values are better.

Metric	Base	PR	PR − Base	Change
Stable RSS	107.9 MiB ± 0.2 MiB	107.0 MiB ± 0.4 MiB	-0.6 MiB ± 0.6 MiB	-0.5%
Physical footprint	61.2 MiB ± 0.8 MiB	61.1 MiB ± 0.2 MiB	-0.3 MiB ± 0.6 MiB	-0.4%
Physical peak	72.9 MiB ± 0.8 MiB	72.9 MiB ± 0.9 MiB	-0.3 MiB ± 0.8 MiB	-0.4%

Interpretation: Differences larger than their measured MAD: Stable RSS -0.6 MiB ± 0.6 MiB.

Extension overhead above pi baseline

Metric	Base overhead	PR overhead	Difference
Stable RSS	+7.0 MiB ± 0.1 MiB	+6.2 MiB ± 1.0 MiB	-0.6 MiB ± 0.6 MiB
Physical footprint	+3.8 MiB ± 1.0 MiB	+3.5 MiB ± 0.5 MiB	-0.3 MiB ± 0.6 MiB
Physical peak	+5.1 MiB ± 0.7 MiB	+4.4 MiB ± 1.2 MiB	-0.3 MiB ± 0.8 MiB

Values are medians of 6 alternating, paired runs. The value after ± is the median absolute deviation (MAD). Each process was sampled 12 times after a 2500 ms warm-up.

Environment: pi 0.84.4, Bun 1.4.4, darwin arm64. RSS comes from ps; physical footprint and peak come from macOS footprint.

This is a comparative signal, not a pass/fail threshold. GitHub-hosted runner noise can affect absolute values.