Update the final candidate status

[suraciii]

What Changed and Why

Update the 0.1.0 roadmap, product release specification, and release designs to
record the final candidate evidence that has now passed on remote master.
Keep the exact tagged proof and release publication open so the repository does
not claim that v0.1.0 exists before a maintainer creates it.

The documented candidate evidence is for 7b65981589151f26f74eb94f29160dccce9a8140.
Because this documentation PR creates a new commit, its exact head and the
resulting remote master will be proved again before this batch is complete.

Related Issue

Tracks #109. This PR does not close the release issue or create a tag.

Specification

• docs/release-0.1.0.md
• design/release.md
• design/release-0.1.0.md, Batch 8e
• ROADMAP.md, Batch 8

The documents now distinguish completed final-candidate proof from the pending
exact-tag proof. No implementation gap is added.

Breaking Changes and Migration

None.

User-Visible Change

NONE

Verification

Exact head be9a3e02aae7b26a86d375f817bd5cca75a3261d, Linux amd64, cold Go test cache:

• make external: pass, 20 external tests, 41.26 seconds. The upgrade proof
  built its source from exact v0.0.5 without a replacement.
• make ci: pass, 77.50 seconds. It ran 638 default tests, 638 race tests,
  20 simulation tests, 19 generator tests, 45 transport tests, 50 shadow
  tests, two resource-bound tests, and three fuzz targets.
• go run ./internal/constraintcheck: pass.
• git diff --check: pass.
• Independent review: no P0, P1, or P2 findings.

Hosted checks for this exact PR head remain required before merge. After merge,
the resulting exact remote master will receive fresh local and hosted proof.

Mutation Test

N/A. Documentation-only change.

Required Checks

• make ci passes locally.
• Old and new tests do not test the same contract.
• Production code does not call time.Now().
• Tests do not use time.Sleep for synchronization.
• Code does not use a mutex to wait across Calls.
• A specification supports the change. It states each implementation gap.

[suraciii]

Exact-head evidence for be9a3e02aae7b26a86d375f817bd5cca75a3261d:

• Cold-cache make external: pass, 20 external tests, 41.26s; upgrade source built from exact v0.0.5 without a replacement.
• Cold-cache make ci: pass, 77.50s; default/race 638 each, sim 20, generator 19, transport 45, shadow 50, resource checks 1+1, fuzz targets 3.
• Hosted run 31499526676: macOS, Windows, Linux stable, Linux 1.25, and aggregate CI all passed.
• Independent diff review: no P0, P1, or P2 findings.

This remains a candidate-status documentation change. It does not create v0.1.0 or claim tagged proof.