Skip to content

docs: re-audit request-for-change rfcs against main - #37153

Merged
KiroCrewCaptain merged 2 commits into
mainfrom
docs/dockeeper-audit-rfc
Nov 14, 2026
Merged

KiroCrewCaptain merged 2 commits into
mainfrom
docs/dockeeper-audit-rfc

Conversation

@iamwhatever

@iamwhatever iamwhatever commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Make every RFC under docs/request-for-change/ say what main does today, so a reader can trust its status and the names it cites.

Many RFCs still said accepted, in-progress or "Nothing built" for work that merged weeks ago. Others cited files, functions and tools that moved or were deleted. The README index had a duplicate row for rfc-browser-live-harness.md, and five rows whose status disagreed with the document's own front matter.

Why it matters

The First Principles review lane reads an RFC's status from the base commit as the record of a product decision. A stale status misleads that lane and every person who reads the index. A path that no longer exists sends a reader to nothing.

Not a goal

What changed (motivation → approach → change)

Each RFC was checked against main at 1d08e941b3: definitions and callers by grep, merged and closed pull requests by gh. The code won every disagreement.

Kind of fix Count
Status moved to partial or implemented, backed by a merged PR 53
Files changed only for a stale path, symbol, tool or PR state 31
README index rows rewritten to match 53
README index rows given a dated correction 24
Duplicate index row removed 1
Gaps left for maintainers (not edited) 30

A status moved only forward, with code on main: accepted, in-progress or partial to partial or implemented. Three moved from partial to implemented (rfc-large-session-bounded-restore.md, rfc-conductor-default-patrol.md, and the rfc-terminal-exit-closes-tab.md index row, whose document already said implemented). Each moved document carries last-audited: 2026-11-14, audited-at: 1d08e941b3, and the merged PR numbers in implementation-prs. The full per-file table is under "Changed files" below.

docs/system-specs/modules/cli.md: the kirocrew memory export/import/migrate row now names each verb with its flags from cli_parser_agent.py (kirocrew memory import <file> [--store <name>] among them), so a search for kirocrew memory import finds it, matching src/kiro_crew/docs/memory-and-learning.md. The row's text is unchanged.

Backwards compatibility

Compatible: docs only; no code, config or schema changes.

Removes nothing: the only deleted line is the second copy of a duplicate README index row; the document it points to keeps its other row.

Tests

Proof: not provable -- docs or config only

  • python3 scripts/docs_lint.py: exit 0 ("All documentation checks passed").
  • ./scripts/docs-lint.sh: exit 0.
  • pytest test/test_docs*.py test/test_pr_quality_gates.py test/test_provider_mirrors.py test/test_cli_doctor.py test/test_crewmate_bundle_export.py: 1009 passed, 50 skipped.

Manual verification

Every one of the 212 .md files under docs/request-for-change/ (205 RFCs, the index, plans/ and assets/) has a line in the audit log: file, verified, errors found, fixed. Each status change was checked with gh pr view and a grep for the shipped symbol.

Related Issues

N/A — routine docs audit.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • 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

Gaps for maintainers

These need a person's decision; the README row for each says so where relevant. Some of these files were still moved forward on code evidence (the entry says so); the open decision is what remains.

Changed files

All 86 changed files and what was wrong
File What was wrong
../system-specs/modules/cli.md kirocrew memory import <file> [--store <name>] was folded into an export/import <file>/migrate label, so it did not read as its own CLI entry
README.md duplicate row rfc-browser-live-harness.md; 53 rows with stale status; 24 rows with stale text (rfc-browser-live-harness and rfc-conversation-trust-follows-founder now read true to main)
assets/autopilot-stage-timeout-evidence.md status note said #14058 "tracks" the cancellation defect; issue closed not-planned 2026-10-04 and #18417 added native cancel at deadline (chat_orchestrator.py refs are pinned 7f1f4fb evidence, left)
plans/2026-08-27-agentcore-identity-gateway.md Task 6 "Modify: src/kiro_crew/security.py" names a file split into the security/ package by #9183 (09c54c7); other missing paths are proposed; header claim (no agentcore_aws/sigv4/inspect) still true
rfc-achievement-feature-discovery.md status accepted but Ship 1 on main (#26512, #26488; Explore behind default-off preview #27449), Ship 2 not built
rfc-agent-config-mirror.md status line listed 4 mirrors; registry.py imports 8 (adds Copilot, DeepSeek, MiMo Code, Strands); KAS still EXTERNAL, mirrors/kas.py & kiro_cli.py are proposed
rfc-agent-plugins-standard.md status in-progress but Phases 1/1b/1c merged via #33621 (agent_plugins/, GET /api/agent-plugins)
rfc-agent-started-turns.md prose status said accepted while fm partial; code has Phase 1+2 (on_agent_turn, dashboard/chat_turn/agent_started.py), label not built
rfc-app-agent-routes.md current-state cites mcp_tools/apps.py, deleted by #33542 (agentTools seam #32660)
rfc-app-contribution-points.md §2.2 "Shipped in #26479" cites apps/link_preview_frame.py, renamed to apps/app_frame.py by #26875 (frame_csp, CAPABILITY_HEADER, _CAPABILITY_TTL_SECS there); draft whose proposals are now on main
rfc-app-sandbox-isolation.md says permissions.mcpTools/cron have no gate; #22965 enforces mcpTools on app agent tools, #19231 checks cron at install
rfc-app-sdk-durable-jobs-and-view-state.md website/src/pages/DevFleetPage.tsx moved to website/src/apps/dev-fleet/DevFleetPage.tsx by #27435 (d5abd13, R099); rest (job_sdk.py, viewState.ts, _RUNS/_ACTIVE_RUNS, dev_fleet_cleanup, no useAppJob) holds
rfc-attended-turn-cap.md status line says #27512 is open and not on main; it merged 2026-11-04
rfc-autopilot-stage-budgets.md status line claimed current main still cancels _run_chat without native session/cancel and that #14058 tracks it; #18417 (merged 2026-10-11) sends native cancel first in _bounded_turn, #14058 closed not planned
rfc-brokered-delivery-confirmation.md status line said #17684 "is open and not on main"; #17684 CLOSED unmerged, same change landed as #27508 (b053285, DeferredTool.on_answered in mcp_shared_calls.py); proposed relayed reason not built
rfc-channel-card-catalog.md status in-progress said Phase 1 is in an open PR; it merged as #35974 (a6d0f72, 2026-11-13): messaging/cards/{spec,render,limits}.py exist and slack/blocks.py builds from Card; Phase 2 (messaging/cards/send.py) absent
rfc-channel-plugin-architecture.md status line's ChannelDescriptor field list omitted restart_credentials (messaging/registry.py); other claims verified (10 descriptors in channels.py, link.py parsers, _resolve_alias, slot stop route)
rfc-chat-core-extraction.md §4.2 table had 8 slices as "in review"/"carried by #8599" that are merged on main: #20018 (ChatEmbed+seed, SideChat, design apps -> sendTurn), #20102 (P3-f Composer root in ChatEmbed), #31093 (P3-d mention.ts), #21152...
rfc-chat-steering-context.md status line said #32179 open / not on main; merged 2026-11-10
rfc-composable-layout-mechanism.md status in-progress but both PRs (#13572, #14062) merged, no open PR, renderer PRs 3-11 not started
rfc-conductor-chat-board-on-tab.md status in-progress but #18608 merged 2026-10-13 and the rule is in agent_parts/spec_prompts.py (prompt moved from agent.py)
rfc-conductor-close-closes-workers.md status accepted + "Nothing is built on main yet", but #30877 MERGED (40e8300, ancestor of main): api_chat_slot_delete calls close_managed in dashboard/chat_api/conductor_close.py; test/test_conductor_close_closes...
rfc-conductor-context-cap.md status line "amendment is built in #29388" read as on main; #29388 is OPEN and main still has Rule A (CONTEXT_CAP_HIGH_TOKENS = 400_000 in dashboard/chat_turn/context_cap.py)
rfc-conductor-default-patrol.md status partial though phase 2 (#19742, merged 2026-10-14) and §10 amendment (#27948) are on main; scripts/patrol_budget.py removed by #33599
rfc-conductor-drift-check.md accepted but #36302 (701e45f) and the amendment's #36618 (551bb52) both merged; goal-conductor/SKILL.md step 5 carries drift check + "re-plan that item now"
rfc-conductor-work-ledger.md status line and §tools say "Five tools ship" on kirocrew-work; mcp_work.py now also carries work_ask_record (CONDUCTOR_TOOLS, #33805) and accept_eval (CHECK_TOOLS, #31849). scripts/ledger.py = security-conduct...
rfc-connections-mcp-layout.md status accepted / "nothing built" but #34353 (one Connections page, MCP Servers page) and #36241 (amendment: /mcp -> Customize tab, McpRedirect) merged
rfc-consolidation-fresh-conversation.md in-progress but #22207 merged 2026-10-19 and fresh_conversation=True / _start_fresh_conversation (llm_helpers_parts/background_turns.py) are on main, used by history_consolidation_pass.py
rfc-crew-log-wake.md "#12781 in flight" but merged (conductor_wake.py present)
rfc-crew-lts-edition.md status accepted with implementation-prs [] but edition.py, LTS gate, sign-in gateway, control-plane policy fetch, LTS baseline in _query_level are on main
rfc-crew-panel-default-tab.md status accepted / "Nothing is built" but #25936 merged; SidePanel.tsx crewFirst behind crew_mode preview
rfc-crewmate-dashboard-upkeep-skill.md status in-progress, implementation-prs [] but #19069 merged (dashboard_upkeep.py, upkeep rows, helpers/); rollout step 3 (v3 Model key) not built
rfc-crewmate-dynamic-dashboard.md status line said the dashboard is still behind the "Dynamic Dashboard" Feature Preview; #19659 (1f6f5af, MERGED 2026-10-13) removed that flag (absent from FEATURE_PREVIEW_FLAGS in config/sections_dashboard.py); das...
rfc-crewmate-goals.md status accepted, implementation-prs [] but #19539 (member_goals.py, goal_reminder.py, goal_ids revision) and #22962 merged
rfc-crewmate-greeting.md cites useMateGreeting / mateGreeting.ts / COLD_AFTER_MS as current; file deleted by #23829 (2026-10-28)
rfc-crewmate-guides-and-mate.md status accepted, "acceptance being requested on the doc PR"; #18326 doc PR, #18301, #18317 merged (mcp_guide.py, first_crewmate.py); mateGreeting.ts / useMeetCrewmatesGate.ts in "checked at 4f9c569" evidence, left
rfc-crewmate-heartbeat.md accepted but #21172 merged 2026-10-21 and the code is on main (member_goals.HEARTBEAT_INTERVALS = (30, 60, 120, 240), goal_reminder.heartbeat_prompt / milestones_card_waiting / gateway_stopping, CrewGoals.ts...
rfc-crewmate-page-ambient-motion.md status line says #22533 open and nothing on main; #22533 merged 2026-10-22, PAGE_LOOK + skill text carry the rule
rfc-crewmate-page-on-goal-wake.md status accepted, "Nothing of it is on main ... both open"; #22533 (eecc217) and #22560 (5c32cf3) MERGED and ancestors of main: dashboard_line/DASHBOARD_BUILD_LINE in dashboard/goal_reminder.py (called from goal_...
rfc-crewmate-profile-settings.md status accepted but #17836 (ce89f96) is on main: approval_mode in config/sections_session.py:397, seed in dashboard/crewmate_approval.py, CrewProfileSettings.tsx, CrewComposer.tsx used by MembersPageThread.tsx. Every §3 rule is built: the profile card sets permission, model and effort, and CrewComposer drops the chips, meter and picker. Later PRs added harness and worker pickers on top; that is extra, not missing, so implemented holds
rfc-crewmate-responsive-send.md status accepted but #20240 merged 2026-10-16 (responsive_chat.py STEER_GRACE_SECONDS = 3.0)
rfc-crewmate-share.md status line "No export route and no owner Import entry exist"; Import #23424 and Export #23223 merged, export moved to .crewjam (#26601); sqlite3.py is an attack example
rfc-crewmate-workers-under-the-crewmate.md status accepted + "Nothing built yet", but migration PRs 2/3/4/6 are on main (#26952 MATE_ROOT_FIELD in crew_mode.py:102, #28869, #28931, #30220 preview in crew_mode.py:585); PR 7 not (chat_folder_file_self still in...
rfc-dashboard-chrome-shell.md both status lines name #16052 as "the implementation"; #16052 CLOSED unmerged 2026-11-01; nothing built (no flush prop on OverlayDrawer.tsx). Pre-scan "missing" paths are files the plan proposes
rfc-delayed-task-brief.md status line said #26938, #26666, #27899 "still open" and the orb still drawn; all three MERGED (2026-11-02/03), #27899 (e049693) deleted WarmBriefOrb.tsx
rfc-deny-session-override.md status accepted, implementation-prs [] but built by #27574 (+ #29427, #33626 agent.deny_override)
rfc-developer-tools-in-settings.md status accepted though #24810 merged 2026-10-28 and code is on main (DeveloperPage.tsx redirects to settings tab developer subs, useDevMode.ts/developerMode.ts gone, app-menu.js visible: true); `electron/app-men...
rfc-drop-goal-item-cap.md status accepted, but #33601 (1c23741) is on main: no DEFAULT_GOAL_ITEM_CAP/MAX_GOAL_ITEM_CAP/item_cap_notice left in src; goal_items_used kept (spec_prompts.py:399); item_cap_exceeded store code kept as...
rfc-followup-start-sends.md status accepted + "Nothing ... built on main", but #25926 (11684ff) is on main: FOLLOWUP_START_SENDS_KEY in website/src/pages/chat/followupStartSends.ts, opt-in send + stay-in-place in useChatPageSlotSwitches.ts ...
rfc-goal-loop-approval-hold.md cites goal-conductor scripts/patrol_budget.py as present; removed by #33599 (1665165, 2026-11-08)
rfc-host-credential-vending.md Problem section cites ACP_BACKEND_ROUTING in agent_sdk/backends.py; defined in agent_sdk/backends_routing.py:463 since #21616 (d9b2e9d); other cited symbols (sandbox_tiers dirs, ENFORCED_ROUTINGS = 3 routings...
rfc-i18n-measurement.md status line says CI runs the overflow gate as the i18n lane of scripts/ci_e2e_parallel.py; #20638 moved it to ci.yml's own i18n-render job. §structural cites scripts/i18n-translate.mjs, real path `website/scri...
rfc-idle-compaction-before-cache-expiry.md status in-progress, "nothing is on main"; Phase 1 merged in #24384 (2026-11-05): idle_compaction in config/loader_builders.py, idle_compaction_delay_secs in agent_sdk/backends_turns.py; Phase 2 Settings UI absen...
rfc-issue-radar-crews.md status line said the two crew ledger tools live in core mcp_tools/apps.py; that file was deleted by #33542 (a974d3e) and the tools moved to the app's app.json agentTools + backend/agent_tools.py by #33156 (7bf6711...
rfc-large-session-bounded-restore.md status partial, but all four delivery parts are merged on main: 1.1 #36259 (8468181), 1.2 #36141 (88307c0), 1.3 #36035 (e618cbe), 1.4 #35960 (c12eb7b)
rfc-mate-landing-once.md status accepted, but its single rollout PR #18904 MERGED (d9557f5, ancestor of main): _stamp_welcome_claimed in dashboard/handlers/members_roster.py, resolveMateLanding in MembersPage.tsx, test/test_members_welcome...
rfc-mate-owned-agent-config.md status line says impl in #21463/#22037 "both open"; both CLOSED, steps 1-3 + migration shipped in #23010 (mate_config.py, mate_config_migration.py, owners.json write guard)
rfc-mcp-routed-server-working-dir.md in-progress "implementation PR is open"; #36277 MERGED 2026-11-14 (538bad7); Proposal C still undecided
rfc-notification-bridge-actions.md banner says "Nothing here is built"; B4a/b/p/c/e and TG/Discord/Webex reply paths merged 2026-11-09..13 (#33749 #33815 #34064 #34474 #34565 #34822 #36050 #36336) behind the bridge_approvals preview; B4f/B4h absent
rfc-notification-bridge.md status line "zero implementation code exists, bridge.py does not exist"; B1 (#32384, #32537), B2 sinks discord/telegram/webex/feishu, B3 Settings UI (#32918) on main; no WeCom sink
rfc-overload-resilience.md taskq/ file map omitted timeline.py (#19942); other "shipped as" homes and not-built items verified; missing test_*.py names are design/plan records
rfc-permanent-patrol.md status accepted, but #33385 (57a1996) is on main: monitor_start arms max_cycles = 0 for a gated loop and pins work-ledger patrols to PATROL_INTERVAL_SECS / no cap / runtime_ceiling_secs() (mcp_tools/control_p...
rfc-perpetual-agent.md status line said implementation "is tracked in" #11582; #11582 CLOSED unmerged 2026-10-29, no Perpetual switch on main
rfc-pipeline-conductor.md head note says UNADVERTISED_AGENTS is in subagent_parts/agent_policy.py; it is defined in subagent_parts/spawn_allowlists.py:45 (agent_policy.py only imports it under TYPE_CHECKING; moved by #35293). Five skill ...
rfc-python-rust-hybrid-backend.md status line says #19627 is an open draft; gh: CLOSED 2026-11-02 unmerged. native/kc_native on main is the redaction gate (#19628), not this RFC's stub
rfc-question-card-pager.md accepted while both parts shipped: #13196 pager and #15250 blocking ask_question (MERGED 2026-10-09, _ask_question_blocking in mcp_tools/control.py)
rfc-recall-process-pool.md status in-progress, impl-prs empty; its one-phase PR #26868 merged 2026-11-01 (memory.recall_pool_processes in config/memory_sections.py)
rfc-remote-computer-use.md Motivation describes the browser panel pull channel (browser-agent-channel.js, command_bus.py, test_browser_command_handler_auth.py) as present; all removed by #26950 (6f13ecb). _computer_use_spec_gate cite...
rfc-remove-piper-tts.md status accepted, "Nothing is removed yet"; now partial: Decisions 1, 2 and 4 shipped in #23189 (piper_runtime.py/piper_worker.py deleted, REMOVED_PROVIDERS = {"piper"} in speech_providers.py); Decision 3 is open, since dev_fleet feature-demo-recording (SKILL.md:92,144, references/deps.py:150-156, narrate.py resolve_provider) and the dev-fleet spec still pick piper under auto
rfc-resumable-subagent-sessions.md subagent_manager/admission.py split into admission/ package (#10521; spawn_impl in admission/gate.py); sync_bridge removed by #10339
rfc-review-loop-and-escaped-bug-reduction.md status accepted + "Nothing is built yet", but all eight decisions are on main: D1 #22698, D2 #22425, D3 #22957, D4 #22971, D5 #22961, D6 #22694, D7 #22973 (pre_push_gate.py), D8 #22702; now implemented
rfc-security-conductor.md cites is_denied in src/kiro_crew/security.py (2 places); file split into a package by #9183 (09c54c7), is_denied now in src/kiro_crew/security/deny_gate.py:461. Other pre-scan misses are false: the scripts ...
rfc-session-template-vs-crewmate.md status accepted but #18342 merged 2026-10-13
rfc-slack-code-channels.md status in-progress, status line says Phase 1 "is in the pull request that carries this document"; #21425 MERGED 2026-10-30, Phase 2 #21410/#22001 merged; the 2026-11-12 Stop-helper-agents amendment is unbuilt (`_sto...
rfc-team-lead-default-agent.md status in-progress though #18128 MERGED 2026-10-11: _install_team_lead_agent in agent_materialization/service_agents.py (called from agent_parts/spec_siblings.py), TEAM_LEAD_AGENT_FILENAME in agent_files.py owned ...
rfc-tool-derived-diff-cards.md status in-progress though #5012 MERGED (db31495, on main) and all in-scope work shipped (isDiffCardItem in TurnBlock.tsx, test_verbosity_config.py); the "unstarted" OutputEvent extension is §3.3 "Explicitly out of ...
rfc-ultra-mode-conductor-takes-defaults.md accepted but its one implementation #31822 MERGED (0760ba5)
rfc-wake-on-stop-signal.md status in-progress though #18854 merged 2026-10-11 (wake_statuses gate in conductor_wake.py, no _wake_bound_conductor left); docs/work-ledger.md is really src/kiro_crew/docs/work-ledger.md
rfc-work-dispatch.md status line "Nothing built" but #25101 merged 2026-10-29 (work_dispatch tool)
rfc-work-fold-forget-closed.md status accepted though its implementation #19068 MERGED (37fcadd, on main): WORK_CLOSED_WINDOW=32 (work_vocab.py) trims closed items with forgotten counts in crew_log/projection_work.py; no item_store_full left i...
rfc-work-ledger-concurrency-cap.md prose status disagreed with front matter; #19338 merged; §7 then shipped in #36312 (DEFAULT_GOAL_CONCURRENCY_CAP = MAX_ITEMS_PER_CONDUCTOR), so status is implemented (re-checked at e51c773a0c)
rfc-work-report-wake-at.md status accepted, "Nothing is built"; #33477 MERGED 2026-11-09 ships wake_at on work_report + storage + staleness deferral; wait_due wake, patrol text, board line not built (no hits in ledger_wake.py code, websit...
rfc-worker-debate.md status accepted + "Nothing built on main yet", but Phase 1 skill change #33783 merged (live gh: MERGED, 92ee944; goal-conductor/references/worker-debate.md); Phase 2 measurement not done
rfc-worker-model-inheritance.md status accepted though #19368 MERGED 2026-10-14: worker_session_defaults (config/sections_agent_fields.py:227) used by dashboard/session_control_parts/session_create.py:679
rfc-workspace-config-evolution.md status line said embeddings.py provides EmbeddingBackend/register_embedding_backend; since #21946 they are defined in embeddings_llama.py / embeddings_shared.py (re-exported); other cited symbols present; plugins/base...
version-compliance-framework.md status line said is_toolbox_install() is unused by production code; doctor_checks/install.py now calls env.is_toolbox_install(); rest holds (update_governance.update_required exists; no set-min-version / --skip-versio...

@iamwhatever
iamwhatever requested review from a team and KiroCrewCaptain as code owners November 14, 2026 08:25
@github-actions github-actions Bot added readiness: checking Automated validation is still running merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Nov 14, 2026
@iamwhatever
iamwhatever force-pushed the docs/dockeeper-audit-rfc branch from 9089ecb to 7d61259 Compare November 14, 2026 08:48
@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Nov 14, 2026
@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision 3689ff21615b853d2c9fafd3799f9d520ea5e789 touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ⏭️ skipped

Revision 3689ff21615b853d2c9fafd3799f9d520ea5e789 ships no reviewable capability (docs, tests or generated files only), so there is nothing to inventory. Does not block merge.

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 3689ff21615b853d2c9fafd3799f9d520ea5e789 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks: this is a docs-only audit. Statuses only move forward, each backed by a merged PR, no RFC leaves draft, and the only line it removes is a duplicate index row.

Notes

  • The First Principles lanes read status: from the base commit (first-principles-review.yml:253). Moving a status from accepted to implemented keeps its weight as a decision record, so this change doesn't loosen what those lanes enforce.

[DESIGN-REVIEWED] 3689ff2

Description read

[DESCRIPTION-READ] 550b1fcd92682411f5524e32b98b6b28c902ee8a66dee150e2d51f1b8a92a6ef

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

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 3689ff21615b853d2c9fafd3799f9d520ea5e789 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 3689ff2

Verdict parsed from the review's SHA-scoped output markers for commit 3689ff21615b853d2c9fafd3799f9d520ea5e789.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 3689ff21615b853d2c9fafd3799f9d520ea5e789 false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 3689ff21615b853d2c9fafd3799f9d520ea5e789 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 3689ff2

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 3689ff21615b853d2c9fafd3799f9d520ea5e789 false-positive: <reason, 20 to 500 characters>

@iamwhatever

Copy link
Copy Markdown
Contributor Author

Rebased onto main to pick up #37171, the fix for the main-red test_tracing_otlp (#37169) that failed Backend Tests (3.12, 2) and Windows (2) here. Windows shard 8's test_two_first_saves_at_once_bind_one_package is the known flake #37114 (recorded there). This PR is docs-only.

@iamwhatever
iamwhatever force-pushed the docs/dockeeper-audit-rfc branch from 7d61259 to 2405973 Compare November 14, 2026 11:47
@KiroCrewCaptain

KiroCrewCaptain commented Nov 14, 2026 •

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: A person should decide

Why: the decider estimates a 7% chance that this change introduces a defect, within the 12% limit for approving on its own, but it changes 912 lines, more than the 400 it may approve on its own.

Docs-only audit of the request-for-change RFCs against main, approved. Every front-matter status change moves forward (accepted/in-progress to partial/implemented, three partial to implemented), and no RFC enters or leaves draft. The First Principles lane counts any status other than draft as a recorded decision (first-principles.md:466-485), so nothing it enforces gets looser. Spot-checks of the cited code all matched the repo: browser_harness/, _close_managed in conductor_close.py, followupStartSends.ts, crewFirst, spawn_impl in admission/gate.py, UNADVERTISED_AGENTS in spawn_allowlists.py:45, REMOVED_PROVIDERS, recall_pool_processes, _ask_question_blocking, mate_config.py and its migration, acp_failover.py, fold_delta.py, messaging/cards/*, pre_push_gate.py, WORK_CLOSED_WINDOW = 32, MATE_ROOT_FIELD, HEARTBEAT_INTERVALS, STEER_GRACE_SECONDS, worker_session_defaults, McpRedirect, isDiffCardItem, _install_team_lead_agent and others. DEFAULT_GOAL_ITEM_CAP and item_cap_notice are gone from src, as the PR says. Tests that read RFC text need exact strings that the PR leaves in place. The cli.md row matches the parsers in cli_parser_agent.py.

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 3689ff2161 · 912 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@iamwhatever
iamwhatever force-pushed the docs/dockeeper-audit-rfc branch from 2405973 to 3943dd4 Compare November 14, 2026 13:19
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Nov 14, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author
  • docs/request-for-change/rfc-chat-core-extraction.md:105 SideChat row has five cells under a four-column header

fixed in 3943dd4: the takeover note moved into the PR cell, matching the #9587 row, so the row has four cells and "merged 2026-10-14" shows in State.
Class swept: every changed table row under docs/request-for-change was checked against its header's cell count; no other row differs.

@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 14, 2026
@iamwhatever
iamwhatever force-pushed the docs/dockeeper-audit-rfc branch from 3943dd4 to 200ff3a Compare November 14, 2026 14:56
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: passed Eligible automated validation passed for the current revision labels Nov 14, 2026
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 14, 2026
@iamwhatever
iamwhatever force-pushed the docs/dockeeper-audit-rfc branch from 5cdd1ad to b0b52ec Compare November 14, 2026 18:42
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Nov 14, 2026

@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.

Thanks for the careful audit. One status in this PR does not match main and needs a fix before merge.

  1. docs/request-for-change/rfc-remove-piper-tts.md:3 (also line 17 and docs/request-for-change/README.md:196). The PR marks this RFC implemented. A reader of the index or the First Principles lane will conclude that Piper is fully removed. But the RFC's own Decision 3 (line 38) and rollout exit 3 (line 55, "the demo-recording skill and the dev-fleet spec no longer name Piper") are not met on main. The dev_fleet feature-demo-recording/SKILL.md:92 still offers --provider piper as what auto picks. Line 144 says "--provider auto prefers piper", and references/deps.py:150-156 still resolves auto to piper. The PR body also lists this file under "Gaps for maintainers (files not changed for status)", which contradicts the diff. Smallest fix: set partial in all three places. Say that Decisions 1, 2 and 4 shipped in #23189 and that Decision 3 (the dev_fleet demo-recording skill) is still open. Keep implementation-prs: [23189].

Optional:

  • The body's gap note for rfc-conversation-trust-follows-founder says no founder trust logic is on main. Main has it in subagent_manager/turn_session.py (_founding_session, resolve_parent_policy), so the suggestion to withdraw that RFC rests on a wrong premise.
  • The kept rfc-browser-live-harness row (README.md:50) still opens with "no browser_harness package exists". The dated correction is attached without a sentence break. Consider adding the break, or putting the correction first.
  • The gap list heading says "files not changed for status", but rfc-crewmate-profile-settings.md is also moved to implemented while its gap entry says the scope widened. Please confirm the widened part is built, or use partial.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: passed Eligible automated validation passed for the current revision labels Nov 14, 2026
zejiangg added 2 commits November 14, 2026 20:05
Bring every RFC's status and current-code references in line with main at 1d08e94, and fix the README index: one row per RFC, status text that matches each document, and notes where a status is a maintainer decision.
Decision 3 of rfc-remove-piper-tts is not on main: the dev_fleet
feature-demo-recording skill and the dev-fleet spec still offer piper
and have auto pick it. The RFC and its index row now read partial.

The rfc-browser-live-harness row no longer opens with a claim that the
browser_harness package is absent, and the
rfc-conversation-trust-follows-founder row records that founder trust
shipped in #30632.
@iamwhatever
iamwhatever force-pushed the docs/dockeeper-audit-rfc branch from b0b52ec to 3689ff2 Compare November 14, 2026 20:10
@iamwhatever

Copy link
Copy Markdown
Contributor Author

Thanks. Fixed in 3689ff2 (rebased onto main first).

Required: rfc-remove-piper-tts is now partial in the front matter (line 3), the status line (line 17) and the README index row. Both say Decisions 1, 2 and 4 shipped in #23189 and Decision 3 is still open. implementation-prs: [23189] is kept. Checked on main: feature-demo-recording/SKILL.md:92 and :144 still offer piper as what auto picks, references/deps.py:150-156 resolves auto to piper, and narrate.py resolve_provider returns "piper" when ready. docs/system-specs/modules/dev-fleet.md:2557 also still lists piper, so the RFC and row name the spec too (rollout exit 3 covers it).

PR body. The piper entry is gone from the gap list, and the "Changed files" row for it now says partial and why.

Class swept: the gap-list heading said "files not changed for status", but rfc-crew-lts-edition, rfc-crewmate-share and rfc-crewmate-profile-settings were moved too. The heading is now "Gaps for maintainers", with a line saying some listed files were still moved forward on code evidence.

Optional points.

  • rfc-conversation-trust-follows-founder: you are right. The gap note now says founder trust is on main from fix: tighten tool approval for continued runs #30632 (merged 2026-11-06), via resolve_parent_policy and _founding_session in subagent_manager/turn_session.py, and it no longer suggests withdrawing. The README row (which this PR had not touched) had the same wrong claim ("never consulted for trust"), so it now carries a dated correction naming fix: tighten tool approval for continued runs #30632. Its status stays in-progress for a maintainer.
  • rfc-browser-live-harness README row: rewritten so it opens with what is true. It is built on main, the harness shipped in feat: add the live browser harness to chat #26950 as src/kiro_crew/browser_harness/, and browser/command_bus.py and browser_cli/ are gone. The stale "no browser_harness package exists" sentence and the missing sentence break are both gone.
  • rfc-crewmate-profile-settings: checked against main. Every §3 rule is built. CrewProfileSettings.tsx sets permission (approval_mode, default trust), model and effort. CrewComposer.tsx drops the agent/model/effort chips, the context meter and the approval picker. crewmate_approval.py seeds the slot. The later harness and worker-model pickers are extra on top, not missing, so implemented holds. Its "scope widened" gap entry is removed, and the Changed-files row says this.

Checks: python3 scripts/docs_lint.py exit 0; ./scripts/docs-lint.sh exit 0; pytest test/test_pr_quality_gates.py test/test_docs_lint_fact_checks.py test/test_docs_lint_owners.py: 240 passed.

@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Nov 14, 2026
@KiroCrewCaptain
KiroCrewCaptain enabled auto-merge (squash) November 14, 2026 22:14

@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.

Thanks for the fixes. On this head the review checked these things. Every added implementation-prs number is a merged commit on main. Every status moves forward only. Each README index row matches its document's front matter. The new memory flags in the cli.md row match cli_parser_agent.py. The strings that the RFC-reading contract tests pin are unchanged.

Earlier review items:

  • rfc-remove-piper-tts status: fixed. It reads partial in the front matter (line 3), the status line (line 17) and README.md:196, and it names Decision 3 as still open.
  • Founder trust gap note: fixed. The note and the README row now cite #30632 and resolve_parent_policy.
  • rfc-browser-live-harness row: fixed. It now opens with what main has.
  • rfc-crewmate-profile-settings: rebutted with evidence that holds. The section 3 rules are built, so implemented stands.

Done well: where the code shows only part of a design, the PR uses partial and names the open phase. Status calls that are out of scope go to a maintainer gap list and are not edited.

The design is signed off under the standing maintainer sign-off for changes by team members.

Optional:

  • README.md:80 (rfc-remote-computer-use) still says the browser pull channel serves a local gateway. #26950 removed that channel. Also, rfc-crewmate-workers-under-the-crewmate.md:19 still calls rfc-mate-owned-agent-config accepted.
  • Some rows still say "Nothing yet" or "Nothing on main" while main has the code. Examples: workflow-library-removal (#27528), acp-backend-failover and fold-delta-push. The dated "Built on main, status left for a maintainer" wording used at README.md:50 would fit them.
  • This PR overlaps #30637 (fully) and #30675 (mostly) in README.md. A "Related:" line would help a maintainer close or trim those PRs after this one lands.

@KiroCrewCaptain
KiroCrewCaptain merged commit 098082d into main Nov 14, 2026
155 of 166 checks passed
@KiroCrewCaptain
KiroCrewCaptain deleted the docs/dockeeper-audit-rfc branch November 14, 2026 22:15
iamwhatever pushed a commit to JiaDe-Wu/KiroCrew that referenced this pull request Nov 17, 2026
…tation PR

The RFC index rows this PR first corrected were re-audited on main by
kirodotdev#37153, and the piper and idle-compaction documents already list their
implementation PRs. One gap is left: rfc-browser-live-harness.md still
has an empty implementation-prs, though kirodotdev#26950 (merged 2026-11-04)
built the harness. This lists it.
iamwhatever pushed a commit to JiaDe-Wu/KiroCrew that referenced this pull request Nov 17, 2026
…k has landed

The RFC index re-audit on main (kirodotdev#37153) fixed most rows this PR first
touched. Two are left: rfc-remote-instance-on-microvm still says the
MicroVM backend half (kirodotdev#17540, merged 2026-10-14) is open, and
rfc-mate-sessions-folder-or-lineage still calls kirodotdev#26952 open. Both rows
now say what merged. rfc-app-contribution-points.md lists kirodotdev#26878 and
rfc-mate-sessions-folder-or-lineage.md lists kirodotdev#26952 in
implementation-prs.
iamwhatever pushed a commit to JiaDe-Wu/KiroCrew that referenced this pull request Nov 18, 2026
…k has landed

The RFC index re-audit on main (kirodotdev#37153) fixed most rows this PR first
touched. Two are left: rfc-remote-instance-on-microvm still says the
MicroVM backend half (kirodotdev#17540, merged 2026-10-14) is open, and
rfc-mate-sessions-folder-or-lineage still calls kirodotdev#26952 open. Both rows
now say what merged. rfc-app-contribution-points.md lists kirodotdev#26878 and
rfc-mate-sessions-folder-or-lineage.md lists kirodotdev#26952 in
implementation-prs.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

readiness: passed Eligible automated validation passed for the current revision

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants