Skip to content

RELEASE_REVIEW.md

Latest commit

 

History

History
116 lines (89 loc) · 5.12 KB

RELEASE_REVIEW.md

File metadata and controls

116 lines (89 loc) · 5.12 KB
Raw

Demonstration Release Review

Review date: 2026-07-21

Frozen build

The evidence manifest, recording, and static states all name this commit. The evidence packaging changes are intentionally separate from the frozen source tag.

Verification results

The full locked suite was run on Windows with Python 3.11.9 and uv 0.11.30:

  • Ruff lint: passed.
  • mypy strict type checking: passed for 29 source files.
  • pytest: 266 passed in 6.63 seconds.
  • Coverage: 1428 statements and 304 branches, 100% reported coverage.
  • Lock validation: uv lock --check passed.
  • Three fresh solguard.demo --skip-paysh processes: all returned VERIFIED.

The three independent fallback runs completed in 6.8929 ms, 9.4199 ms, and 7.6594 ms. Each produced a compound-drain BLOCK, NOT_SIGNED, unchanged post-baseline balance, and a successful recovery payment.

On this Windows checkout, core.autocrlf=true converts existing tracked Python files to CRLF, so a local ruff format --check . reports 22 existing files as different from the configured LF form. The exact frozen commit's required Ubuntu CI formatting check passed. No source file was reformatted merely to conceal that host-specific difference.

Local decision benchmark

Environment:

  • Microsoft Windows 11 Home 10.0.26200
  • 12th Gen Intel Core i5-12500H, 16 logical processors
  • Python 3.11.9
  • uv 0.11.30

Method: 500 sequential PaymentGateway.process calls with a deterministic clock, valid known-recipient requests, all local security controls, and in-memory simulated settlement. Network activity and Pay.sh settlement are excluded.

Metric Measured result
Allowed decisions 500
Other decisions 0
Minimum 0.0992 ms
Median 0.1222 ms
Mean 0.1515196 ms
p95 0.2561 ms
Maximum 0.5068 ms

Reproduce from the frozen source:

uv run --with pillow --with qrcode --with imageio --with imageio-ffmpeg `
  python scripts/generate_release_evidence.py --output evidence --iterations 500 `
  --source-commit v0.1.0-demo

Results vary by host and run. Do not compare these local figures to network settlement latency.

External sandbox observation

One Pay.sh sandbox demonstration returned SETTLED for a 0.01 USDC request and then completed the same local attack/recovery sequence. Its total command duration was 5433.5796 ms. This is a single captured sandbox observation, not a latency benchmark, mainnet payment, or production availability claim.

Dependency and repository security review

  • The project declares no runtime Python dependencies.
  • uv.lock was exported with all development groups; pip-audit checked the 13 applicable third-party packages and reported no known vulnerabilities on 2026-07-21.
  • GitHub secret scanning, secret-scanning push protection, and Dependabot security updates are enabled.
  • Safe API queries returned zero open secret-scanning alerts and zero open Dependabot alerts at review time.
  • A local detect-secrets scan of the changed non-binary files reported only the public Pay.sh sandbox recipient address and evidence digest/hash fields. They were reviewed as non-secret evidence values.
  • main and develop require the Verify Python 3.11 check, current branches, pull requests, linear history, resolved conversations, and administrator enforcement. Force pushes and branch deletion are disabled.
  • GitHub code scanning has no analysis configured. Required approving review count is zero and signed commits are not enforced by branch protection. Those are explicit hardening gaps, not passing controls.

No secret value was printed or retained during this review.

Evidence integrity

evidence/manifest.json records the exact source commit, labels, artifact sizes, and SHA-256 hashes. Independent media inspection confirmed 750 frames at 10 fps, a 75-second duration, H.264 video, yuv420p pixel format, and 1440 x 900 resolution. The QR image decoded to the repository URL.

Known limitations and release decision

This is an evidence-backed demonstration release, not a production release. Remaining gaps include process-local replay and authorization state, no tenant isolation or authenticated policy administration, no hardened key custody, no high-availability design, no code-scanning workflow, no independent assessment, and only one external sandbox adapter. x402 is not implemented.

The tagged build is suitable for the documented sandbox/offline demonstration. It is not approved for real funds, production traffic, or security guarantees beyond the verified scenarios.