Stability measurement: baseline, sweep experiment (reverted), roadmap 7-9 - #8
Merged
Conversation
One row per changed contract — names, shared-object identity, generated strings other code compares against, documented attributes — with the consumer search recorded per row. Targets the enumeration misses the 5x2 baseline measured (required_methods flip and Rule.methods introspection at 2/5 while blockers sat at 5/5). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This reverts commit 7ff8362.
…igned - benchmarks/stability-2026-08.md: 5x2 cold-run baseline (blockers 5/5, zero noise), the contract-sweep experiment, and why it was reverted — tail variance is the seven-finding cap reallocating slots, not an enumeration defect. - ROADMAP: PR/MR comment surface, subagent auto mode, deep mode as an open question. - Output rules: auto mode's 'fix the BLOCKERs' is the after-the-review ask — the review itself still contains no patches. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Measures finding stability across runs, documents the result, and records the roadmap items agreed in review.
required_methodsflip) was discovered and verified during the measurement.7ff8362) was tested with the same protocol and reverted (b0599b2): it lifted its target key (2/5 → 4/5) but paid by dropping others; runs render ~4 cards against a 6-key union, so the cap reallocates slots rather than expanding coverage. Requests lost consistency (Jaccard 1.00 → 0.80, one verdict flip). Full numbers in the report.Non-Python smoke (not committed, rendered for review)
One cold run each on sirupsen/logrus#1574 (Go) and fastify/fastify#6965 (JS): comment markers came out in the language's own syntax (
// ←), behavior demos used realgo run/injectoutput inconsoleblocks instead of a fabricated Python REPL, tables matched cards. Both reviews also had teeth (logrus: the fix lands in a function the logging path no longer calls; fastify: the advisory's sibling entry point still bypasses the not-foundpreHandler).Do not merge without an explicit go-ahead.
🤖 Generated with Claude Code