Skip to content

Cross-client audit run from the full property set #91

Closed as not planned
Closed as not planned

Description

@grandchildrice

M2 · track:2-C · pipeline — target 8/15 (step 5) · depends on the full-EIP cross-check issue · blocks the reproduction and report issues · part of #86

Background

With the property vocabulary fixed, run the audit stage (02c → 03 → 04) against every target client. SPECA runs the left stage once and the right stage per implementation, so this is a fan-out over clients with properties held constant. The output is the per-client confirmed / disputed findings that 2-C reports on.

Tasks

  1. Define the client matrix — the production implementations (the 10 Fusaka clients are the reference set: go-ethereum, Nethermind, Prysm, Lighthouse, Lodestar, Nimbus, Grandine, Reth, and the others). Confirm the current list and pin commits.
  2. For each client, prepare outputs/BUG_BOUNTY_SCOPE.json (trust model, severity rubric, deployment_context.target_share) and outputs/TARGET_INFO.json (repo + commit + language).
  3. Run 02c → 03 → 04 per client via scripts/run_phase.py (or the CI workflows), using resume so a mid-run failure doesn't re-spend tokens.
  4. Collect 04_PARTIAL_*.json per client and triage verdicts (CONFIRMED_VULNERABILITY / CONFIRMED_POTENTIAL / DISPUTED_FP / DOWNGRADED / NEEDS_MANUAL_REVIEW).
  5. Confirm each finding carries its spec_reference (already a 04 field) for the report.
  6. Capture a per-client run manifest: commit, property-set version, cost, counts.

Done when

  • Every client in the matrix has a completed 04 run with findings JSON.
  • A triaged, per-client finding list exists with spec references attached.
  • Runs reproduce from the pinned commits + property-set version.

Review checklist

  • Each client has both config files with a defensible target_share.
  • Findings carry spec_reference; NEEDS_MANUAL_REVIEW items are flagged, not dropped.
  • The manifest records exactly what ran, so the report issue is auditable.
  • Per-client BUG_BOUNTY_SCOPE trust levels match the client's real input surfaces — not copy-pasted loosely.

Out of scope

Reproducing findings on a devnet. Disclosure decisions and the written report.

Activity

  1. added
    M2Milestone 2 grant work
    track:2-CDeliverable 2-C: cumulative bug & fix report
    pipelineFull pipeline runs
    on Jun 30, 2026
  2. grandchildrice commented on Jul 22, 2026

    @grandchildrice
    CollaboratorAuthor

    2026-07-22 M2 再構成のためクローズ。 3トラック(a=gasper / b=EL / c=vuln-dataset)× 4フェーズの新構成に作り直しました → トラッカー #86。

    この issue の引き継ぎ先: トラック別の全クライアント監査 a-3 #148 / b-3 #149 / c-3 #150 に分割。

    作業内容・議論は新 issue とトラッカーに移行済みです。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

M2Milestone 2 grant workpipelineFull pipeline runstrack:2-CDeliverable 2-C: cumulative bug & fix report

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions