[ScriptedAlchemy]

Choice and cause

Implements the removal option in #4341. No signing account, certificate, submission, download, or signature-verification stage is wired into either release channel or the shared package action. The checked-in SignPath portal XML and indefinite rollout wording implied that Windows executables would be signed even though published executables remain unsigned.

Changes

• Remove the unused .signpath/artifact-configuration.xml portal configuration.
• Replace the signing rollout claim and incorrect #79 citation in SECURITY.md with the current unsigned state. This is a one-line edit because docs(security): clarify daemon authority and outbound operations #4348 also touches that file.
• Clarify in the user guide that stable archives have build-provenance attestations and that Windows executables are unsigned. Checksums and provenance attestations do not provide an executable code signature.
• Leave archive and MCPB build, packaging, checksum, attestation, and publication behavior unchanged.

Choosing signing instead would require provisioning a signing account and certificate, adding a required submission/download and publisher/chain verification step before both archive and MCPB packaging in stable and beta, failing publication on signing errors after a defined cutoff, and verifying the downloaded published artifacts on Windows. That wiring is outside this removal PR.

Verification

Before the change, rg --hidden -n -i 'signpath|authenticode' .github .signpath scripts tests docs SECURITY.md README.md found the unused XML and SECURITY rollout claim; release workflows and shared package action had no signing stage. A red unit test is not applicable to this workflow/documentation cleanup.

After the change, rg --hidden -l -i 'signpath|authenticode' across the source tree returned no files; the only Windows signing statements now say the executables are unsigned. git diff --check passed. No workflow YAML changed, so actionlint is not applicable.

• bash tests/release_safety_test.sh, passed (exit 0)
• bash tests/release_drift_check_test.sh, passed (exit 0)
• bash tests/dashboard_bundle_check_test.sh, passed (exit 0)
• bash tests/release_pr_integrity_test.sh, passed (exit 0)
• scripts/check-release-drift.sh --release-version 1.0.0-beta.89 printed release versions are aligned: 1.0.0-beta.89

Bazel/clippy are not needed for a workflow and documentation cleanup with no code or workflow changes.

Fixes #4341

---

[changeset-bot]

⚠️ No Changeset found

Latest commit: f2823bb

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no bugs or issues to report.

[ScriptedAlchemy]

Pushed dcf78fedef54e08ce9dab33c499359b371bce11a for the inherited hook journey failure in CI run 39519779223. Both failed attempts were legitimately accepted for replay; the test assumed that its single retry would satisfy the unchanged 25 ms acknowledgement deadline.

The journey now retries the same immutable Stop event only after a completed accepted_for_replay, retryable: true receipt and typed replay Drained readiness. It uses one existing 180-second deadline across attempts. Exact guidance, single emission, duplicate suppression, and decision-engine assertions remain; production behavior and budgets are unchanged.

Pinned rustfmt, diff checks, and commitlint pass. Native execution is delegated to the new CI run.

This fixes the shared hook fixture. The separate fact_detail_keeps_a_superseded_fact_readable_by_id HTTP timeout in the previous run is not addressed by this change and remains for the new CI run to diagnose.