Skip to content

docs(rfc): re-check every RFC status against main - #39948

Merged
KiroCrewCaptain merged 3 commits into
mainfrom
docs/sweep-r1b-rfc-status-2
Nov 20, 2026
Merged

KiroCrewCaptain merged 3 commits into
mainfrom
docs/sweep-r1b-rfc-status-2

Conversation

@bolichen97

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Make the status of every RFC, and its README Index row, match the code on main after a full re-check.

The earlier passes checked mostly symbol names. A full re-check found RFCs whose status lags shipped work, README rows that disagree with their own front matter (one uses in-review, which is not a status value), partial RFCs that understate what is built, and draft RFCs that say "nothing built" while code is on main.

Why it matters

The First Principles review reads RFC status from the base branch as the product decision record. A wrong status lets a change through, or blocks one, on a false premise.

Not a goal

What changed (motivation → approach → change)

All 227 RFCs were re-checked against main: every implementation-prs number read live, README row vs front matter vs body status line, a merged-PR search per RFC name, and definition+caller checks for each candidate deliverable with the maintainer decision looked up on the doc and implementation PRs. Status moves past draft only with a recorded maintainer decision.

For the reviewer: rfc-browser-live-harness moves to implemented on the owner decision on #26950, which also says its status "moves to accepted after a maintainer design review". If that review is still owed, it should stay draft. rfc-delayed-task-brief should list the two greeting RFCs in supersedes (left for its owner).

Per-RFC table and evidence
  • Worktree: /home/bolichen/.kiro/crew/scratch/29751d83/wt-rfc2, branch docs/sweep-r1b-rfc-status-2, base origin/main 0c5d8bd5f3 (2026-11-20).
  • Commits: 2d635e295b, 77bdd207e8, 3d17187c88 (not pushed). ./scripts/docs-lint.sh exited 0 before each commit.
  • Scope edited: front matter status, last-audited, audited-at, implementation-prs, superseded-by; the in-body status line; README Index rows. No RFC body was rewritten and no file was deleted. CHANGELOG.md was not touched.

How each RFC was checked

  1. Front matter of all 227 rfc-*.md files was extracted. Every implementation-prs number (398 PRs) was read live through the GitHub GraphQL API: 376 merged, 12 closed, 10 open.
  2. Each RFC's README row status was compared with its front matter, and its body status line with both.
  3. gh search prs --merged "<rfc-name>" ran for every RFC. Merged non-docs PRs that name an RFC but are missing from its implementation-prs were listed and triaged.
  4. Every backticked path and identifier in each README row was checked against the files on main (65 rows had a missing token; each was triaged, and most were proposals or deliberately deleted code).
  5. For each candidate, the deliverable was checked in code (definition and caller), and the maintainer decision was looked up on the doc PR and implementation PRs: accept comments, sign-off reviews, and /ai-review override first-principles comments. Maintainers per MAINTAINERS.md: bolichen97, iamwhatever, buluoray, CrysisDeu.

Status changes

RFC Old status New status Evidence
rfc-acp-backend-failover draft partial Phase 1 is acp_failover.py, read by dashboard/messaging_api/spawn.py (#26123, merged). Decision: the approving review of doc PR #27581 says "A maintainer has signed off on this RFC's design". Phase 2 (runtime failover) is not built.
rfc-app-contribution-points draft partial The document said it "flips to accepted when the document merges"; it merged in #26859 with the owner's §12/§13 decisions. On main: apps/app_frame.py messageActions (#26878), shared frame (#26875, #26899), selectionActions (#27700), composerActions apps/composer_action_frame.py (#27718), linkOpen apps/link_open_frame.py (#27737), agentTools apps/manifest_agent_tools.py (#32660). The §8 roadmap is not built.
rfc-subagent-backend-selection draft partial Decision: bolichen97's comment on #18150, "Maintainer decision (bolichen97): yes, I want this feature". Phase 1 (#18150) has _requested_backend refusals. Phase 2 is not built: api_spawn_continue has no backend refusal.
rfc-crew-owned-oauth draft partial Decision: bolichen97's review on #29517 (2026-11-04), "Maintainer decision (security sign-off): the maintainer signs the direction". Code is connections/crew_oauth*.py behind the crew_oauth flag in FEATURE_PREVIEW_FLAGS (default off). The phase 3 default-on flip is not done.
rfc-browser-live-harness draft implemented #26950 (merged 2026-11-04) shipped the whole stack, src/kiro_crew/browser_harness/, and removed browser/command_bus.py and browser_cli/. Decision: bolichen97's owner decision on #26950 (issuecomment-6569794150) plus the matching /ai-review override first-principles comments.
rfc-resource-control accepted partial Accepted when #39344 merged. §4 (re-wakes while the host stays tight, per-worker growth) shipped in #39604 (merged 2026-11-20). §3 (CPU) and §5 (wake filter) have no code.
rfc-crewmate-greeting accepted superseded (by rfc-delayed-task-brief.md) Shipped in #17797 and #17822. The warm branch was replaced by the warm task brief (WarmBriefCard.tsx). #23829 removed the cold Recap on the owner's decision. useMateGreeting, COLD_AFTER_MS and mateGreeting.ts are gone.
rfc-crewmate-greeting-as-message accepted superseded (by rfc-delayed-task-brief.md) Shipped in #18651 (now in implementation-prs). MateResumeCard and MateWelcomeCard are gone after #23829 and the warm brief replaced them.
rfc-memory-injection draft partial The document records "A maintainer has signed R1 only". R1's authority order shipped in #35222. Shadow-first code for the unsigned stages (#34260, #35194, #35211, #35231, #35460, #35461, #35789) is on main, default off. The status line says those stages are not decided.
rfc-slack-code-channels partial partial (row and status line corrected) The row said the 2026-11-12 and 2026-11-16 amendments were not built. All of them are now on main: stop_helpers.py (#38243), #38112, #38240, actions.py and comments.py (#22370), tabs.py (#38998), #38181. The §3 "Stop: What happens next" brake is not built; it is in open #38882.
rfc-otel-tracing partial partial (row and status line corrected) The document said "Phases 2 to 9 are not built". Phases 2 (#35792), 3 (#36153, #36717) and 4 (#37005, #37158) and the 5a round traces (#36833) are merged, in src/kiro_crew/tracing/. Phase 2b (SigV4) and phases 6 to 9 are not built.

Stale README rows fixed (front matter was already right)

RFC Row said Now Evidence
rfc-agent-self-rename in-progress implemented #33736 merged; _apply_chat_rename in dashboard/session_directive_apply_slot.py
rfc-bot-subfolder-person-delete accepted implemented #38808; delete_empty_bot_subfolder, called from api_chat_folder_delete
rfc-conductor-context-cap partial implemented #22718, #27024, #29388 merged
rfc-delayed-task-brief partial implemented all 17 listed PRs merged, the last #38307
rfc-fold-recorded-edit-cards accepted partial _snapshot_write_target (#25650) is on main; no file_change_recorded under src/
rfc-host-memory-guard accepted implemented #39270, #39289, #39755; pressure_reclaim (called inside session_memory_guard.py), apply_guard_report
rfc-jev-task-executor accepted partial only #16014 (rollout step 2) is on main
rfc-kas-session-head-in-agent-prompt accepted implemented #37108; session_head.py
rfc-mate-sessions-folder-or-lineage accepted partial path B PRs merged; chat_folder_file_self still granted
rfc-remote-instance-on-microvm draft implemented #17540; cloud/microvm/
rfc-rm-root-row-scratch-entries draft implemented #36536; security/rm_floor.py
rfc-session-tag-change-event in-review (not a status value) draft #32962 and #36132 merged, #7669 closed; no acceptance recorded
rfc-slack-owner-claim accepted implemented #39086 (the body status word was also changed accepted→implemented)
rfc-slack-shared-access in-progress implemented #30760; slack.shared_access
rfc-update-architecture in-progress partial update_capability.py, wheel_engine.py, update_stepup.py on main
rfc-work-ledger-person-wait-hold partial implemented #18109 and #38307 merged

Row path fixes with the status unchanged:

  • rfc-issue-radar-crews: the tool registration is apps/builtins/issue_radar/backend/agent_tools.py, not mcp_tools/apps.py.
  • rfc-crewmate-goals: the path is member-records/member-goals/, not trust/member-goals/.

Draft or accepted kept, with the status line, implementation-prs and row corrected

RFC Status Correction
rfc-app-sdk-browser draft It said "Nothing in this document is built". Steps 1a, 1b and 1c are on main behind app_browser_sdk (#39018, #39022, #39020). bolichen97's comment on #38322 says it is "not a sign-off".
rfc-mcp-resource-subscriptions draft It said "Nothing is built". #24713 (Phase 1 test) and #34124 (Phases 2 to 4, merged 2026-11-20) are on main. No acceptance and no override on #34124.
rfc-crew-log-telemetry-export draft It said "Nothing is built". S1 (#21342, metrics/crew_log_export.py, installed from server_runtime/maintenance.py) and S2 (#21308) are on main. Both were merged by chenmingwei23 while iamwhatever's HOLD reviews waited on design agreement.
rfc-app-backend-adoption-provenance draft It named #13650 as "the implementation"; that PR closed unmerged. #35334 shipped Removal 1 as an opt-in setting.
rfc-conductor-remote-dispatch draft implementation-prs was empty; 13 merged wave PRs were added.
rfc-efficient-local-inference draft Added #21763 (Phase 2's power_source()) to implementation-prs.
rfc-session-keyed-scratch draft Added #30062 (Phase 1 rejoin). The row now says Phase 1 is on main.

Audit-field-only bumps (last-audited 2026-11-20, audited-at 0c5d8bd5f3)

Needs a maintainer ruling

Not done (left for a later pass)

  • Deep per-phase code reads were done only for the RFCs above. The other ~150 partial and implemented RFCs were checked through PR state, row/front-matter consistency, the merged-PR mention search and the row token check, and no issue was found. They were not re-read phase by phase, and their audit fields were not bumped.
  • The merged-PR mention search lists PRs missing from many RFCs' implementation-prs (mostly refactors). It is in $KIROCREW_SCRATCH/r1b/missing.txt, and implementation-prs was completed only where a status depended on it.
  • The plans/*.md files carry no status front matter or status line, so they were not changed.
  • The README audit-history paragraph was not extended with this pass.
  • version-compliance-framework.md has an Index row but is not an rfc-*.md; it was not checked.

Risk: decisions recorded only in conductor chats or off GitHub were not visible, so any draft above may already have an off-repo acceptance.

Backwards compatibility

Compatible: documentation only; no code, config or API changes.

Removes nothing: only doc wording changes; no file, key or function is removed.

Tests

Proof: not provable -- docs or config only

./scripts/docs-lint.sh exits 0.

Manual verification

Each status was checked against the symbols and PRs named in the table above.

Real-machine verification

N/A -- docs only: the diff changes Markdown front matter and index rows and runs nothing.

Related Issues

Follows #39909. Open questions for a maintainer are in the "Needs a maintainer ruling" part of the table above.

Checklist

  • A Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...); PR Hygiene does not count commits
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

Contribution License Agreement

Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.

@bolichen97
bolichen97 requested review from a team and KiroCrewCaptain as code owners November 20, 2026 11:32
@bolichen97
bolichen97 requested a review from buluoray November 20, 2026 11:32
@bolichen97
bolichen97 enabled auto-merge (squash) November 20, 2026 11:32
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 20, 2026
@github-actions

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ⏭️ skipped

Revision fcfda806b48d62e5527b491f7afb55c555b707a8 ships no reviewable capability (docs, tests or generated files only), so there is nothing to inventory. Does not block merge.

@github-actions

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision fcfda806b48d62e5527b491f7afb55c555b707a8 touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of fcfda806b48d62e5527b491f7afb55c555b707a8 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Docs-only status and Index-row corrections under docs/request-for-change/; nothing to block.

Notes

  • The PR moves rfc-browser-live-harness from draft straight to implemented, but the decision it cites on feat: add the live browser harness to chat #26950 says the RFC "moves to accepted after a maintainer design review". If that review hasn't happened, the move breaks the PR's own rule: "No status is moved to accepted without a recorded maintainer decision". Keep it draft unless a maintainer confirms the review took place.

[DESIGN-REVIEWED] fcfda80

Description read

[DESCRIPTION-READ] a8ccde44d1b0b9993578314ab018ea6ebe3b356ceab9d09b8241ae0194ab7596

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed fcfda806b48d62e5527b491f7afb55c555b707a8 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] fcfda80

Verdict parsed from the review's SHA-scoped output markers for commit fcfda806b48d62e5527b491f7afb55c555b707a8.

False positive or not applicable? A repository writer can comment:
/ai-review override fable fcfda806b48d62e5527b491f7afb55c555b707a8 false-positive: <reason, 20 to 500 characters>

@github-actions

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of fcfda806b48d62e5527b491f7afb55c555b707a8 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] fcfda80

False positive or not applicable? A repository writer can comment:
/ai-review override gpt fcfda806b48d62e5527b491f7afb55c555b707a8 false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 20, 2026
@KiroCrewCaptain

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: A person should decide

Why: the decider raised 1 possible problem and estimates a 55% chance that this change introduces a defect; a person should check it.

This is a docs-only sweep of RFC statuses and README Index rows, and most of it is well evidenced and consistent. One change contradicts the PR's own rule: rfc-browser-live-harness moves from draft to implemented on an owner decision that explicitly defers acceptance to a maintainer design review, and no such review is recorded. Similar RFCs in the same PR (code on main, no acceptance) stay draft. The First Principles lane reads implemented from the base branch as a recorded product decision, so this would count an unaccepted design (including the no-flag removal of playwright-cli) as decided. The author and the Design Review lane both flagged it. It should stay draft unless a maintainer confirms the review.

Things to check (possible problems the decider raised; at this confidence many turn out to be fine):

  • README 'Status vocabulary' says the First Principles lane reads implemented from base as a recorded product-shape decision and that a PR flipping the field proposes, not records, a decision. Promoting this RFC to implemented on a decision that explicitly defers acceptance to a later maintainer d (docs/request-for-change/rfc-browser-live-harness.md, medium)

Tier: 2, held for a person: design-doc-gate: D.

An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit fcfda806b4 · 390 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@KiroCrewCaptain KiroCrewCaptain left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I reviewed head fcfda80. It changes only RFC front matter, the in-body status lines and the README Index rows under docs/request-for-change. The docs lint passes on this head. Every status value is in the documented vocabulary, every Index row now matches its document's front matter, and every superseded-by target exists. No test reads these fields, and the review gate that does reads them from the base branch, so nothing here changes behaviour. One non-blocking note for a maintainer: the browser live harness RFC now reads implemented, while the owner decision it cites deferred acceptance to a later design review, so please confirm that status or revert it to draft.

@KiroCrewCaptain
KiroCrewCaptain merged commit 5dfde5a into main Nov 20, 2026
98 checks passed
@KiroCrewCaptain
KiroCrewCaptain deleted the docs/sweep-r1b-rfc-status-2 branch November 20, 2026 12:35
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 20, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants