Repository navigation
docs: align 40 docs with main after the week of 2026-11-14 - #37935
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Docs-only change that brings the specs back in line with main; no code, contract or persisted data moves, so nothing blocks. [DESIGN-REVIEWED] 85d8701 Description read[DESCRIPTION-READ] bcb5e5b48c78a08fe0544c053ea09f8be5541b9e82c7a3987192677b7640702a 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 |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS This PR changes only docs, and each edit brings a doc in line with code already on main. No item changes a product default. What this change shipsInventory (10 items) — 10 justifiedIntent: make the docs say what main's code does after a week of merges. This is a FIX (docs drift).
[FIRST-PRINCIPLES-REVIEWED] 85d8701 Description read[DESCRIPTION-READ] 09663bf9d42c3ea7050f8d6c5c20ed6b185b8b261012917f9e11a4ad2cb3112f That sha256 names the PR description this verdict read. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
|
b434f3c to
87fe067
Compare
87fe067 to
e340f60
Compare
The memory spec described record-list filters (applies, source, status, never_recalled), a status-widened refresh and the unforget and set_tier bulk operations. None of them is in the records or bulk handlers on main; they belong to the open #37498. The config spec named SessionAllocation, whose class is SessionAllocationService. The runtime-ownership paragraph on a restart rejoin is dropped, since #37826 removed join_on_tree and _join_own_tree.
e340f60 to
85d8701
Compare
Code change approver recommendation: ApproveWhy: the decider estimates a 10% chance that this change introduces a defect, within the 12% limit for approving on its own, and every check passes. This PR changes docs only, to bring them in line with main. I spot-checked about 30 cited symbols, constants and limits and all of them match the code. That covers the Windows deny-gate helpers, rm_floor._RM_WORDS, the import size limits, BOT_MAX_WORKERS_MAX=32, browser change-report budgets, clone identity, monitor terminal wake, the user-rate-limit text, privacy notice version 3, channels_enabled default false, the otlp base dependency at the same pin, and the /v1/crew/owner/bots route. Removing the memory-record filters and the unforget/set_tier operations is correct: no source under src/ mentions them. The RFC status moves are backed by merged PRs, as in the earlier audit wave; nothing touches docs/decisions/ and no trust boundary moves. The only gap is that acp-client.md leaves out MimocodeAcpClient. That is incomplete documentation with no behavioural effect, so it does not block. 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 |
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
This review checked the doc claims on this head against the code on main. All changed .py paths resolve. The new client classes, SessionAllocationService and the FEATURE_PREVIEW_FLAGS list all match. So do the import size limits (2 GiB, 100, 64 GiB, 1 GiB), the Windows deny-gate helpers, the policy validate exit behaviour and the meetings CSP. The RFC status moves cite PRs that are merged (#37609, #37164, #34104). The PR touches no code, tests or baselines.
Well done: the memory-skills-hooks.md cut removes record filters and bulk operations that main's records handler refuses. It also points to #37498, the open PR where that code lives.
The design is signed off under the standing maintainer sign-off for changes by team members.
Optional:
- docs/system-specs/modules/config.md:1649 and src/kiro_crew/docs/snapshot-and-restore.md:321 say every import refusal names the biggest folders. Only the 64 GiB ceiling and free-disk refusals do. The per-file and zip-bomb ratio messages do not.
- docs/system-specs/modules/acp-client.md:67 lists four AcpClient subclasses.
MimocodeAcpClient(acp/client_mimocode.py) is already in this PR's base and is missing. - docs/request-for-change/README.md:62 says #34104 built
GET /v1/crew/owner/bots, then says the only versioned route isPOST /v1/chat/completionsat4e605b0d2b. Thelast-auditedandaudited-atfields in the three RFCs whose status moved still show the older audit.
Problem / Motivation
Goal: Make the docs say what the code on main does after the 304 commits that landed between
b81cf5a3ac(2026-11-14) and5abc52be3a.Those commits renamed classes, moved owners, changed defaults and added CLI flags, config keys and preview names. About forty docs still described the old code or left the new parts out.
Why it matters
People and agents read these docs before they touch the code. AGENTS.md sends every session to the owning spec first, so a stale owner table or a missing flag sends them the wrong way.
Not a goal
docs/decisions/(byte-immutable). No drift was found there.What changed (motivation → approach → change)
Each behaviour commit in the range was read against the docs that name its module, tool, flag, key or route. A doc was fixed only where the current code says otherwise.
cli.mdremote-control status/on/offorworkspace list/delete; theevalrow lacked--baseline,--max-drop,--junitacp-client.md,side.md,model-selection.md,agent-host-contract.md,harness-gap-analysis.md,harness-onboarding.md,error-handling.mdClaudeAcpClient,GooseAcpClient,OpencodeAcpClient,DeepseekAcpClient) and their owner modules were missing; moved methods and tripwires pointed at old homes; new rate-limit wording was undocumentedconfig.md,configuration.md,snapshot-and-restore.mdconfig.mdnamedSessionAllocationforSessionAllocationService; preview names list lackedbridge_approvals,secrets_vault,worker_grill; no row fordashboard.channels_enabled; import size limits and rebuildable folders left outsecurity.md,governance.md,meetings.md,auto-improvement.mdpolicy validateon an unlistable profiles dir; sketch frame CSP now hash-pinned; clone identity pinmonitor-architecture.md,monitor-loops.md,crew-log-projection.md,session-control.md,crew-members.md,app-notifications.mdneeds_yourule; per-Bot worker cap; Goals tab credits card;bot_limits_heldnotemetrics.md,learn-cron-dashboard.md,dashboard.md,api-reference.md,browser.md,knowledge.md,kas-auth.md,remote-crew.md,feature-map/README.md[GRILL:]marker, browser change report, eval tool assertions, knowledge ignore caps, SSO start URL, Shared Crews cardinstall.md,docker.md,windows-install.md,e2e-gate.mdotlpextra, newer container image on About, installer clearing too-deep filesmemory-skills-hooks.mdapplies,source,status,never_recalled), a status-widened refresh and theunforget/set_tierbulk operations that the records and bulk handlers on main do not accept (they are in the open #37498)rfc-reroute-first-ask-policy.md,rfc-work-ledger-resource-pressure.md,rfc-crew-api.md,rfc-crewmate-workers-under-the-crewmate.md,request-for-change/README.mdimplemented;rfc-crew-apimoves topartialsince #34104 serves the owner roster; the crewmate RFC no longer cites a retired previewBackwards compatibility
Compatible: docs only; nothing that works today stops working.
Removes nothing: the diff only corrects or adds prose and table rows.
Tests
Proof: not provable -- docs or config only
python3 scripts/docs_lint.pyexits 0; report-only dead identifiers went from 605 to 601.bash scripts/docs-lint.shexits 0.scripts/check_brand_name.pypasses on the added lines.Manual verification
Each changed claim was checked against the code at
5abc52be3a.Related Issues
N/A — weekly docs-vs-code check, no tracking issue.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license.