docs(skills): dual-walk boolean axes + sweep-delta evidence step (post-#5072 R4/R5) - #5100
Conversation
…phase-rs#5072 R4/R5) The phase-rs#5086 dual-walk paragraph predates phase-rs#5072's review rounds 4-5. Adds the three decision-bearing boolean axes (reads_member_bound, reads_event_live, writes_event_object) that drive profiles_conflict's same-event discriminators, and the post-classification full-DB sweep step with the predicate-class vs genuine-exact-set distinction (auto-absorb vs completeness-asserted explicit add). Symbols grep-verified against main at 8f47155. Assisted-by: ClaudeCode:claude-fable-5
There was a problem hiding this comment.
Code Review
This pull request updates the documentation in .claude/skills/add-engine-variant/SKILL.md to detail the classification process for the read/write conflict profiler. Specifically, it adds instructions regarding the three decision-bearing boolean axes (reads_member_bound, reads_event_live, and writes_event_object) and explains how to run and handle the full-DB parity sweep. There are no review comments, so no feedback is provided.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
matthewevans
left a comment
There was a problem hiding this comment.
Thanks for the follow-up. I need to hold this one for direct maintainer handling because the diff edits .claude/skills/add-engine-variant/SKILL.md.
Skill and agent instruction files control future automation behavior, so the PR-review sweep does not accept or auto-merge external contributor changes in those paths. This is a maintainer-gate issue rather than a judgment on the substance of the wording.
Please leave this for Matt to review directly or rework the change through a maintainer-owned branch.
🤖 AI text below 🤖
Summary
Follow-up to #5086: the dual-walk paragraph it added to the
add-engine-variantwalker checklist predates #5072's review rounds 4–5, which added real classification obligations. This documents (1) the three decision-bearing boolean axes (reads_member_bound,reads_event_live,writes_event_object) that driveprofiles_conflict's same-event discriminators — mis-setting any is the exact #5072 R3-HIGH-1 / R4 failure mode; and (2) the post-classification full-DB sweep evidence step: predicate-keyed conservative classes absorb new corpus members automatically, but the genuine exact-sets are completeness-asserted and require an explicit per-card-CR-evidenced addition (never a floor bump).Docs-only companion in the #4905/#5086 pattern.
Files changed
.claude/skills/add-engine-variant/SKILL.mdCR references
CR 603.3b / 603.10a / 608.2h (context in the documented axes; no engine code touched).
Track
Developer
LLM
Model: claude-fable-5
Thinking: high
Tier: Frontier
Verification
Docs-only change — no engine/parser code touched:
8f47155a9:se_member_bound_class×16,se_event_object_class×15,SAME_EVENT_MEMBER_BOUND_GENUINE×5,SAME_EVENT_EVENT_OBJECT_GENUINE×4,reads_and_writes_event_object×12,event_object_present×27,ordering_parity_sweep×2.cargo fmt --all— no-op (no Rust files).Scope Expansion
None.
Validation Failures
None.
CI Failures
None.
Related: #5072 (the R4/R5 rounds being documented), #5086 (the paragraph being extended), #5073 (why the sweep step is manual). Per the
.claude/skills/**policy this is direct-maintainer-review.🤖 Generated with Claude Code