Repository navigation
docs: encode review corrections and fold the resource-ID RFC into main's design - #815
Draft
behinddwalls wants to merge 1 commit into
Draft
behinddwalls wants to merge 1 commit into
behinddwalls wants to merge 1 commit into
Conversation
behinddwalls
force-pushed
the
preetam/agents-learnings
branch
3 times, most recently
from
October 8, 2026 02:57
551fda2 to
69470a5
Compare
behinddwalls
changed the title
behinddwalls
force-pushed
the
preetam/agents-learnings
branch
4 times, most recently
from
October 8, 2026 03:31
69f6550 to
c60ab85
Compare
…n's design ## Summary ### Why? Recent sessions kept getting the same review corrections: speculative packages and adapters, homegrown wrappers, host code in the library, framework layouts, non-Bazel tooling, re-encoded IDs, internal names in a public repo, separate CI workflows, and RFCs that describe discussion rather than main. Autoreview also never ran its architecture or correctness lenses on TypeScript, because `.ts`/`.tsx` files classified as `other`. One such RFC already exists: queue-scoped decimal IDs shipped (#770, #804), but the decision still lived in a standalone "current vs proposed" RFC. ### What? - `AGENTS.md`: new **Design Defaults** section, aligned with the web UI RFC (#810); CI credential and workflow rules; naming, RFCs-describe-main, and concise-docs bullets; ask before splitting a large change into a stack; local run steps in the Test Plan. - Autoreview `review-scope.py`: new `web` kind (`.ts`, `.tsx`, `.js`, `.css`; `*.test.*` as tests), and `internal-reference` and `new-workflow` smells. - Autoreview dispatch: the parent runs a planner, one lens reviewer per assignment in parallel, and a consolidator, because subagents cannot start their own. Committed targets are reviewed in a clean detached worktree so local edits never skip checks, and every readable changed file must appear in some reviewer's Read list or the report marks it unreviewed. Coverage is counted from Read/Partial/Unread buckets rather than judged, and every finding must be re-opened at its cited line before it is admitted, with a recheck log in the report. - Autoreview `review.md` / `lenses.md`: route `web` to all code lenses; treat unneeded added surface as should-fix; add checks for each design default and for RFCs that diverge from main; sweep commit messages for internal names; skip data-migration findings for OSS deployments. - RFCs: add a **Resource IDs** section to `doc/rfc/submitqueue/workflow.md`, delete `doc/rfc/scoped-resource-ids.md` and its index entry, fix stale ID text in the history and Stovepipe workflow RFCs, delete the superseded SubmitQueue `list-api.md` proposal (shipped as `status-list-api.md`), and describe the implemented Stovepipe List API as current rather than proposed. ## Test Plan ✅ `bazel test //.agents/skills/autoreview/scripts:all` ✅ `bazel test //tool/docsite:site_test` (strict link check) ✅ `make fmt` leaves the tree unchanged
behinddwalls
force-pushed
the
preetam/agents-learnings
branch
from
October 8, 2026 03:43
c60ab85 to
cb26534
Compare
This branch has not been deployed
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.
Summary
Why?
Recent sessions kept getting the same review corrections: speculative packages and adapters, homegrown wrappers, host code in the library, framework layouts, non-Bazel tooling, re-encoded IDs, internal names in a public repo, separate CI workflows, and RFCs that describe discussion rather than main. Autoreview also never ran its architecture or correctness lenses on TypeScript, because
.ts/.tsxfiles classified asother. One such RFC already exists: queue-scoped decimal IDs shipped (#770, #804), but the decision still lived in a standalone "current vs proposed" RFC.What?
AGENTS.md: new Design Defaults section, aligned with the web UI RFC (feat(web): add a read-only SubmitQueue web UI as a mountable module #810); CI credential and workflow rules; naming, RFCs-describe-main, and concise-docs bullets; ask before splitting a large change into a stack; local run steps in the Test Plan.review-scope.py: newwebkind (.ts,.tsx,.js,.css;*.test.*as tests), andinternal-referenceandnew-workflowsmells.review.md/lenses.md: routewebto all code lenses; treat unneeded added surface as should-fix; add checks for each design default and for RFCs that diverge from main; sweep commit messages for internal names; skip data-migration findings for OSS deployments.doc/rfc/submitqueue/workflow.md, deletedoc/rfc/scoped-resource-ids.mdand its index entry, fix stale ID text in the history and Stovepipe workflow RFCs, delete the superseded SubmitQueuelist-api.mdproposal (shipped asstatus-list-api.md), and describe the implemented Stovepipe List API as current rather than proposed.Test Plan
✅
bazel test //.agents/skills/autoreview/scripts:all✅
bazel test //tool/docsite:site_test(strict link check)✅
make fmtleaves the tree unchanged