firewall: let the samplers faculty name the autolens lane tiers - #174
Merged
Conversation
PyAutoBrain's SamplerSurface now scans the findings maturation lane's experiment (autolens_workspace_developer/searches_minimal) and mature (autolens_profiling) tiers, closing the "Surface gap, filed" note the faculty carried in its own AGENTS.md. Naming either repo in _samplers.py / samplers.sh trips the tenant firewall — both files are pinned to their current three tokens, and a NEW instance fact in an allowlisted file is drift. Verified by running the check before the edit: it flagged exactly those two files and nothing else. This grows two existing entries rather than adding new ones, which is the form the surrounding comment permits: a new entry means a new file an adopting fork must rewrite, and there is none here. The Brain's new tests/test_samplers_surface.py takes its tier labels from module constants precisely so it needs no entry. Also records PyAutoMind in the prompt's `Repos:` header — it listed only PyAutoBrain, which is not implementable as written. Brain side: PyAutoBrain@claude/automind-task-planning-8wk4ag (94c18ed). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WSCA1h6cSKMNwFmWEsxuCL
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Mind-side companion to PyAutoBrain#216, which teaches the samplers faculty's
SamplerSurfaceto scan the findings maturation lane's experiment (autolens_workspace_developer/searches_minimal) and mature (autolens_profiling) tiers — closing the "Surface gap, filed" note the faculty carried in its ownAGENTS.md.Prompt:
draft/feature/pyautobrain/samplers_surface_autolens_tiers.mdWhy this repo is involved at all
Naming either repo in
_samplers.py/samplers.shtrips the tenant firewall. Both files are pinned inFIREWALL_ALLOWLISTto their current three tokens, and a new instance fact in an allowlisted file is drift.Verified rather than assumed — running
scripts/repos_sync.py --checkbefore the edit flagged exactly those two files and nothing else:Both are clear after this change.
Why growing entries is the right form here
This grows two existing entries rather than adding new ones, which is what the surrounding comment permits — a new entry means a new file an adopting fork must rewrite, and there is none here. Brain's new
tests/test_samplers_surface.pytakes its tier labels from module constants precisely so it needs no entry of its own.Also
Records
PyAutoMindin the prompt'sRepos:header. It listed onlyPyAutoBrain, which is not implementable as written — the firewall makes this a two-repo change — and adds a short note in the body explaining why, so the constraint is visible to whoever reads the prompt next.Noted, not fixed
scripts/repos_sync.py --checkalso reports one pre-existing mismatch unrelated to this change:PyAutoBrain/tests/test_worktree_conflict_guard.pynames 7 repos and has no allowlist entry. It landed in PyAutoBrain13d222c(#215). The check is not wired into any CI workflow — it needs all sibling checkouts — which is how it slipped through. Left alone deliberately: adding a new file entry is the consequential form the code warns against, and that call belongs with the author of that test.🤖 Generated with Claude Code
https://claude.ai/code/session_01WSCA1h6cSKMNwFmWEsxuCL
Generated by Claude Code