[kixelated]

Problem

moq-dev/rind#505 finished the last child of the m2 speed epic, quest/m2/speed/README.md, which left the epic as an [M] quest whose only remaining work was checking its own goal. That goal (every algorithm on every path as fast as the fastest permissive source, or the gap recorded with its reason) is already enforced on the committed data: CI's admin/bench-report render --check refuses any gap over 10% that has no benchmarks/gaps.toml entry. Re-measuring is the job of quest/m3/bench-report-refresh.md, which requires m2, so the epic was only holding that refresh back.

Approach

• Delete the speed epic and its link from m2's README.
• The m3 refresh is now the audit. Its Plan carries the epic's settled bar (the sources, tiers, and sizes a gap is judged against) and says how new gaps become quests.
• Drop the epic's link from the Known gaps text in admin/bench-report and regenerate BENCHMARKS.md.
• Point the bench-report test's quest fixture at the permanent /quest/m3/README.md.
• Move the epic's note about RSA A/B runs to a comment on bench.yml's no-ifma row, where a reader of that row needs it.

Impact

• No public API or wire change. BENCHMARKS.md's Known gaps intro no longer links an epic.

Decisions

Settled with the maintainer on 2026-11-09 through /quest-complete and /quest-plan:

1. What to do with the speed epic now that its last child is done: ✅ audit, then decide (rejected: /quest-delete it now, keep as an [M] quest). The plan's first round then found the overlap above.
2. Which goal this plan pursues: ✅ delete the epic now, with the m3 refresh as the audit (rejected: keep the epic as an [M] audit, which duplicates the refresh on 2-day-old data; move the refresh into m2, which reverses the 2026-11-03 decision that the refresh runs at the start of the next milestone).
3. Where speed quests the refresh finds go: ✅ the refresh lists them as suggested follow-ups in its PR body, which /quest-merge plans into the milestone the maintainer picks (rejected: recreate a speed epic in m3, which would leave an empty epic waiting).

Checks

• quest gate just check origin/main passed: the bench-report tests, render --check, quest, gh, shell, and nix checks.
• quest check --base origin/main: 148 documents ok.

Models

Size: XS
Implementer: Claude Code, claude-opus-5-5, high effort

(Written by Claude Opus 5.5 with high)

🤖 Generated with Claude Code

Review

Rewritten by quest review after each round.

• Reviewer: Codex, gpt-6.1-sol, low effort
• Session: 01a1bf0d-9eb0-7151-963b-d1f0fb116dd2, round 1
• Head: 83d53ae
• Base: origin/main at 8a5b5be
• Score: 10/10, The changes consistently remove the completed speed epic and preserve its relevant guidance elsewhere, with no behavioral or security defects identified by code inspection.
• Findings: none

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 11 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: dc0de875-c099-403a-ace1-af4abe3699ca

Commits

Reviewing files that changed from the base of the PR and between 8a5b5be and 83d53ae.

Files selected for processing: 8

• .github/workflows/bench.yml
• BENCHMARKS.md
• admin/bench-report
• admin/bench-report-test
• benchmarks/gaps.toml
• quest/m2/README.md
• quest/m2/speed/README.md
• quest/m3/bench-report-refresh.md

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn

[pullfrog]

✅ No new issues found.

Reviewed changes

• Speed epic deleted — quest/m2/speed/README.md is removed along with its Required bullet and the "speed and armv7" wording in quest/m2/README.md's Goal, which now matches its two remaining children (api, adoption).
• Settled decisions carried into the m3 refresh — quest/m3/bench-report-refresh.md gains the epic's bar (fastest permissive Rust-usable source, every nightly-bench tier, the 32 B/1200 B/1350 B/16 KB sizes plus 64 B/8 KB for hashes, copyleft excluded) with its 2026-11-01 attribution, and the new 2026-11-09 triage decision (gaps needing a quest become suggested follow-ups in the refresh PR's body).
• Report text and regeneration — the Known gaps intro drops the epic link in admin/bench-report and the regenerated BENCHMARKS.md; benchmarks/gaps.toml's header comment is generalized to match.
• Test fixture repointed — admin/bench-report-test's gap fixture now links the permanent /quest/m3/README.md, whose actual title m3: deferred matches the new expectation (gap_table reads the title from the repo file).
• RSA note relocated — the epic's note about A/B runs reading bench.yml's no-ifma row is condensed into a comment on that row itself.

Verified: render --check passes (byte-identical regeneration), all 17 bench-report-test tests pass, no entry in gaps.toml referenced the deleted epic, and a whole-tree grep finds no stale m2/speed or "Fastest at every algorithm" references. The epic's handshakes/bulk-transfer scope is inherited by the refresh since admin/bench-report itself times those rows, and the "real Pi 4 stays unmeasured" caveat is substantively covered by the CPU-tiers intro's "GitHub's runners … with CPU features denied". The body's account of PR #505 finishing the last child and of the 2026-11-09 maintainer decisions is consistent with the repo state.

  ^{｜ View workflow run ｜ Using glm-5.3 ｜ 𝕏}