Skip to content

feat(crew): give each crewmate its own config, a conductor and a worker - #23010

Merged
CrysisDeu merged 1 commit into
mainfrom
feat/crewmate-2
Oct 23, 2026
Merged

CrysisDeu merged 1 commit into
mainfrom
feat/crewmate-2

Conversation

@buluoray

@buluoray buluoray commented Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Make each crewmate the conductor of its own work: it gets its own config file, chats on a gateway-assembled conductor, and hands every task to its own worker, which runs with that crewmate's kit and permissions.

Today a crewmate is a pinned DM thread bound to a shared agent template. Its chat both talks to the user and does the work, so a long job ties up the thread, its tools are whatever the template happens to mount, and there is no place to give one crewmate an extra server or skill without forking the template. The accepted RFCs (#22443) split the two roles: a conductor that answers, plans and dispatches, and a worker that does the file and command work.

Why it matters

A crewmate is meant to keep at a goal for days. With one session doing everything, the thread goes silent during work, the user cannot see or take over what is running, and every crewmate on the same template shares one tool set. Splitting conductor and worker keeps the chat free, puts the work in sessions the user can open, and lets each crewmate carry its own kit without touching anyone else's.

Not a goal

What changed (motivation → approach → change)

Goal → approach: the RFCs accepted in #22443 (docs/request-for-change/rfc-mate-owned-agent-config.md and the conductor/worker RFC, both accepted on main) set the split. This PR is the Crewmate 2.0 core of the trial branch #22037, re-applied on current main with everything outside it left out (list below). It is one commit.

  • No switch. Every crewmate runs this way; there is no dashboard.crewmate_v2 key, toggle or off path. New crewmates are 2.0 from creation, and existing ones migrate once, one way.
  • Each crewmate owns its config. mate_config writes ~/.kiro/mates/agents/<id>.json, holding only what that crewmate adds. The startup migration (mate_config_migration.migrate_existing_crewmates) moves a crewmate off a shared template or a crew-<hex> copy, skips and logs one whose worker would differ, repoints its own DM thread records to the conductor, and posts a one-time upgrade notice in its chat. Nothing moves a crewmate back. The ledger entry records the crewmate's member_id, so a key deleted and made again starts a fresh entry.
  • Conductor. kirocrew-mate-conductor is assembled by the gateway: the platform servers only, the default template's built-in tools (the shell and file writer included), and no spawn / task / workflow runners. Sub-agent spawning is the one thing it loses (conductor_spawn_refusal); it may run a quick command or a small edit itself and hands longer work to its worker. It keeps its notes with a new save_notes tool. This PR amends rfc-mate-owned-agent-config.md (§5.0, §6 step 1, §10 question 7) to record that decision, the KAS remap below, and the conductor's scripted cron; the earlier read-only, no-shell conductor is superseded.
  • Worker. Every session_create from a crewmate opens its own kirocrew-mate-worker: the gateway base, plus the template kit copied in the crew editor's new Advanced > Worker's extra kit section, plus capabilities the user added. It runs on the crewmate's member backend and only on a backend that takes a per-session spec (ACP_BACKENDS_MATE_WORKER_KIT = {kas}); elsewhere the hand-off is refused with worker_kit_harness_unsupported. When a worker ends a turn, its report wakes the crewmate (dashboard/mate_thread_wakes.py). A crewmate's scheduled agent turn runs on its worker too, not the conductor: at each fire mate_worker.cron_worker_execution turns the job's member capture into the worker's record, so the turn carries the crewmate's kit, servers, permissions, store and identity under the job's own cron: key. A 1.0 job (template_id: kirocrew with a member id) takes the same path with no data migration. A scripted cron_add from the conductor is dispatched like any caller's (test_cron_conductor_scripted.py); there is no conductor gate for it, by the design owner's decision recorded in the RFC amendment. When that turn's result changed (the cron's own last_posted_hash dedupe), the crewmate's thread gets the usual [Worker report] wake (mate_thread_wakes.note_cron_result), recognised by the job's attested member binding and its cron: key; an identical result wakes nobody, including a job the worker itself created.
  • Permissions. The Profile's Permissions setting reaches every worker of that crewmate (crewmate_approval.py).
  • A worker's gap. A worker reports a missing tool as MISSING CAPABILITY:. request_capabilities is refused on that wake turn (not_user_turn), so the wake tells the conductor to name the gap to the user and raise main's card on the user's next turn. A worker started after the card applies has it.
  • Review fixes. An approval removed in a crewmate's Approvals stays off even when the worker base grants it (mate_approval_revocations.py), and Approvals still lists it switched off so it can be turned back on. Approvals also lists the worker base's own grants, which a migrated file does not hold, so each can be switched off. The worker-kit preview lists every tool a copy turns off. A worker-kit copy keeps the template's excludedTools, and a template whose permissions or toolsSettings limits the worker cannot carry is refused (a create from it keeps the crewmate on the template). A worker-kit apply is bound to the file AND the result its preview showed, so a template edited since asks for a fresh preview. A create whose template has the crew's own name is refused (template_name_in_use). The migration does not move an agent whose permissions deny/ask rules or toolsSettings its worker would lose, and its repoint is a compare-and-set on the template it planned from, so a switch made meanwhile wins. A worker whose crewmate file, revoked-approvals record or execution record cannot be read, or whose crewmate's own file is absent or shadowed, is refused, not run on the bare base; revocations are read even for a crewmate with no file. The migration writes its ledger intent before a repoint and marks the crewmate moved after, so a crash in between still moves its old threads. Revocations are written before a capabilities save, re-grants only after it. A crewmate may not take a name a user-level agent declares, and a crewmate file another instance sharing the Kiro home wrote is never replaced (mates/owners.json). An unreadable crewmate file no longer reads as removed, so it cannot reset the default agent. save_notes replaces the file through a temporary and refuses a symlinked leaf. A fresh install keeps plain chats on the configured template; only the first crewmate runs on the conductor.
  • Built From reads the worker. The roster's Built From names the worker kit's template, and a kit read error shows as an error rather than as the default template.
  • KAS. A crewmate chat needs the dispatch tools that only kas carries (main's own doc for member_acp_backend in config/sections_agent.py: a value resolving to kiro leaves member threads as plain chat, no dispatch tools), so a stored kiro member backend now runs as kas, by the design owner's decision. The remap is logged once at load with the stored and the used value; config.json keeps kiro, and the remedy "set member_acp_backend to kiro" is gone because it no longer changes anything.
  • Prompts in one file. The conductor's member section, the worker's principles, and the goal-setup, welcome, worker-report and upgrade wakes now live in src/kiro_crew/crewmate_prompts.json. scripts/crewmate_prompt_preview.py assembles them offline from fixtures. The goal-setup kickoff now hands research to a worker session instead of spawn_run.
  • Spec reads through agent_sdk. agent_sdk/agent_specs.py is the one seam the crewmate modules read agent specs through, so none of them imports kiro_crew.acp.

Re-applied on main's split modules: members_parts/member_sessions.py (worker backend routing), handlers/members_roster.py, handlers/core_config_schema.py, chat_turn/turn_admission.py, and CrewEditorNextDialog.tsx, which replaces the old editor that main removed. main's #21430 told a crewmate without file tools not to keep its briefing. Here a crewmate with the kirocrew-mate tools can still write it through save_notes, and #21430's text moved into the prompt file unchanged.

Moved to follow-up pull requests, opened after this one merges: Dynamic capabilities (the per-crewmate preview, find_capabilities, named additions, the ask card, the hand-off capability note and the same-worker reload), and each crewmate's own git worktree.

Left out of #22037, and the small pieces kept:

Left out What stayed, and why
Heartbeat 1f35c185c5, 5898b34b2b Nothing. main's heartbeat (#21172) is used as is, with its hand-off line ending the wake so the worker's report wakes the thread. The trial's quiet-end folding of hidden-turn text (heartbeat_quiet_end.py) and its prompt keys are dropped.
Dashboard 12db47d80d, 2d66cf544b, 07a715166e, a0c7cc7855 Nothing. The dashboard.* prompt keys are gone, and the conductor's crew-log line no longer names the trial-only folds.
Duplicate card 800956d11b On main via #22513.
Milestones by number eb2b648386 The parts 2.0 needs that had been folded into that commit: the worker-report wake (mate_thread_wakes.py), mate_welcome.app_is_stopping, the per-turn permission hook, and the prompt preview script. key_result_met, milestones_line and the milestones.* keys are dropped.

Backwards compatibility

Compatible: an upgrade migrates every crewmate, once. That move is the product-shape change the accepted RFCs record (#22443). It is one way: no switch turns it off. A shared template and any crew-<hex> copy are never deleted.

Reader: src/kiro_crew/dashboard/handlers/members_roster.py:api_members -- crew page -- a binding the migration left on a legacy crew-<hex> copy (mate_worker.is_legacy_binding) is not listed twice; test/test_legacy_binding_roster.py
Reader: src/kiro_crew/agent_materialization/service_agents.py:_remove_user_level_copy -- release -- deletes ~/.kiro/agents/kirocrew-mate-{conductor,worker}.json only when name and description are the generated ones, and keeps any other file with a warning; test/test_mate_service_specs.py
Reader: src/kiro_crew/mate_worker.py:MATE_CONDUCTOR_EXCLUDED_TOOLS -- chat -- a crewmate's chat loses only the spawn runners (its built-ins are the default template's); work goes through session_create to its worker; test/test_mate_conductor_agent.py, test/test_mate_worker_dispatch.py
Reader: src/kiro_crew/dashboard/handlers/cron_jobs.py:api_cron_tools -- cron -- a scripted cron_add from a crewmate's conductor is dispatched like any caller's (the conductor's scripted-cron refusal is removed); test/test_cron_conductor_scripted.py
Reader: src/kiro_crew/mate_config_migration.py:migrate_existing_crewmates -- crew page, chat, cron -- agents.<name>.kiro_agent and the crewmate's own DM thread template_id are rewritten, once; test/test_mate_config_migration.py
Reader: src/kiro_crew/config/loader_builders.py (member backend) -- chat -- a stored kiro is read as kas; config.json keeps kiro; test/test_acp_backend_kas.py

Tests

New suites, each pinning one behaviour:

  • test_mate_owned_config.py, test_mate_config_migration.py, test_mate_migration_real_kirocrew.py, test_mate_upgrade_notice.py: the own-file write, the migration (what it carries, what it skips, a template restriction it would lose, idempotence, the thread repoint, a fresh ledger entry for a key made again) and the notice being posted once.
  • test_acp_backend_kas.py: a stored kiro member backend runs as kas and the remap is logged once; test_api_agents_create_template.py: a template or user-level agent of the crew's own name is refused; test_mate_worker_kit.py: an apply after the template changed is refused, and adopting a user-level agent's name is refused; test_mate_worker.py: a revoked approval stays off, and an unreadable crewmate file or revocation record refuses the worker; test_mate_config_migration.py: a template switched during the pass is not overwritten; test_guide_server.py: a fresh install keeps plain chats off the conductor; test_mate_owned_config.py: another instance's crewmate file is not overwritten; test_config_loader_default_agent.py: an unreadable crewmate file is not removal evidence. test_cron_crewmate_worker.py: a crewmate's cron turn runs on its worker with its MCP servers, a 1.0 record moves onto the worker once, and a non-crewmate cron is unchanged. test_cron_crewmate_report.py: a changed crewmate cron result wakes the conductor once, an identical one does not, and a non-crewmate cron never does; a worker-created job reports too. test_mate_worker_kit.py: a template with uncarried limits is refused on copy and on create, and its exclusions are carried. test_mate_owned_config.py: a switched-off approval stays listed and turns back on, and a base grant missing from the file switches off and on; test_mate_worker_kit.py also checks the preview lists an added exclusion.
  • test_mate_conductor_agent.py, test_mate_conductor_every_create_path.py, test_mate_service_specs.py, test_mate_service_names.py, test_cron_conductor_scripted.py: the conductor's tool set (built-ins kept, spawn tools excluded), every create path landing on it, where the service specs live, and a scripted cron from it being accepted.
  • test_mate_worker.py, test_mate_worker_dispatch.py, test_mate_worker_kit.py, test_mate_worker_disabled_servers.py, test_mate_platform_server_override.py, test_mate_thread_wakes.py: the worker merge (additions only, switched-off tools withheld), the worker spec read by member id, backend routing and refusal, the kit copy, and the worker-report wake.
  • test_crewmate_worker_approval.py: Permissions reaching workers.
  • test_mate_capabilities_redaction.py, test_mate_capabilities_worker_summary.py: the in-place editor's redaction and the worker counts on Built From.
  • test_legacy_binding_roster.py, test_member_notes.py: the roster and save_notes.
  • Frontend: WorkerKitPane.test.tsx, builtFromName.test.ts, workerKitLabels.test.ts, KiroCrewAgentsPage.rosterLabels.test.tsx (including a kit read error), plus updated CrewProfileSettings.test.tsx and CrewEditorNextDialog.test.tsx.

Local run on this head: black, isort, flake8, mypy --platform linux, agent-SDK boundary, subprocess encoding, line cap, harness parity, docs-lint, comment-history, the prompt-preview golden, tsc, i18n:check, and the touched Python and Vitest suites, and test/integration/test_governance.py. test/e2e/test_private_workflow_memory.py now expects the nested spawn from a crewmate's workflow to be refused (its chat runs on the conductor).

Manual verification

On the trial branch (#22037) this slice ran through five pod soak rounds and a real-install switch. On this head, in a pod: a crewmate bound to kirocrew was migrated at startup (ledger entry with its member_id), posted the one-time notice, handed "list the files in your workspace" to its worker with session_create, and the worker's report woke it with the answer. The pod checks of the ask card and the same-worker reload move with the Dynamic capabilities PR.

Screenshots / video

A migrated crewmate's chat: the upgrade notice, the hand-off to its worker and the worker's report (light):

Conductor and worker chat, light

The same crewmate's Profile, Permissions reaching its workers, no Dynamic capabilities row (light):

Profile, light

Related Issues

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

Contribution License Agreement

Contributed under the project's Apache-2.0 license.

@buluoray
buluoray requested review from a team and KiroCrewCaptain as code owners October 22, 2026 03:46
@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 Oct 22, 2026
@github-actions

github-actions Bot commented Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

Sensitive Change Review

Sensitive files changed:

  • src/kiro_crew/security_posture_sinks_gateway.py

  • @buluoray posted a complete reasoning template for 1934364.

@github-actions github-actions Bot removed the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 22, 2026
@github-actions

github-actions Bot commented Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — 🔴 BLOCK

Premise-level review of 1934364508c03793bff44348f8c7737d271311ef — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: BLOCK

product-shape change without accepted RFC: this PR reverses the conductor tool set that the base RFC marks as settled, and it also remaps a stored kiro member backend to KAS. Both decisions are recorded only in an RFC amendment that the PR itself adds.

Notes

  • Item 4 — undeclared: hand-off was available on main for claude/codex/opencode/goose members, and the base RFC never mentions removing it. Hunk: raise SessionControlError(backend_reason, code="worker_kit_harness_unsupported"). Subtraction: limit own_worker forcing to ACP_BACKENDS_MATE_WORKER_KIT members, so other backends keep main's dispatch.

What this change ships

Inventory (12 items, capped at 10) — 7 justified

Intent: each crewmate gets its own config, a conductor chat and its own worker that does the work (ADDITION).

  1. A crewmate's chat now has a shell and file writers, but no spawn/task/workflow — unjustified: reverses settled §5.0
  2. Each crewmate gets its own config file; existing crewmates migrate once, one way, and get an upgrade notice — justified
  3. Every hand-off from a crewmate opens that crewmate's own worker, whatever agent is named — justified
  4. Hand-off refused on non-KAS member backends — undeclared
  5. A stored kiro member backend now silently runs KAS — unjustified: in-diff amendment only
  6. A crewmate's scheduled agent turns run on its worker; a changed result wakes the thread — justified
  7. A worker's report wakes the crewmate thread — justified
  8. Crew editor: new Advanced > Worker's extra kit section — justified
  9. Profile Permissions and Approvals revocations reach workers — justified
  10. Roster "Built From" names the worker kit's template — justified

Blockers

product-shape change without accepted RFC — conductor tool set and kiro→KAS remap (items 1, 5)
Trigger: the base rfc-mate-owned-agent-config.md (accepted) says "no shell (execute_bash is not mounted)… Settled: read-only… Recorded here so the question is not reopened in review." The diff deletes that text and adds "Amended 2026-10-23 in #23010… keeps… execute_bash, fs_write and code… A stored kiro member backend runs as kas." The sections_agent.py help text drops "'kiro' selects kiro-cli."
Outcome: every crewmate chat gets a shell and file writer by default, and an operator's explicit kiro choice is overridden. The base commit records no decision for either change, only this PR's proposal.
Clears when: an RFC covering the read/write conductor and the kiro→KAS member remap is merged non-draft on the base (then push or rebase), or a maintainer posts /ai-review override first-principles 1934364508c03793bff44348f8c7737d271311ef: <decision>.

[FIRST-PRINCIPLES-REVIEWED] 1934364

Description read

[DESCRIPTION-READ] 077155e0afd5b2fbaf78a4bb9b8f130a28aa594a890227b541df624f0d59481a

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Oct 22, 2026
@github-actions

github-actions Bot commented Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — 🔴 changes requested (blocking)

GPT 6.1 found at least one blocking issue that must be resolved before merging 1934364508c03793bff44348f8c7737d271311ef. 2 of 3 blocking finding(s) are security-class and were withheld from adjudication, so the blocking verdict stands.

This comment is updated in place on each push.

BLOCKING -- src/kiro_crew/mate_config.py:184 -- Concurrent creates bypass the foreign-file ownership check

owners = _read_owners(); _refuse_foreign_file(mate_id, owners)
Trigger: Two instances with separate data homes sharing the mate directory concurrently create the same key through adopt_own_file → write_mate_spec; both read the ownership record before either writes it. The dashboard lock is process-local.
Outcome: Both creates succeed, but the second write replaces the first instance’s config, losing its worker kit. Concurrent ownership-record writes also discard entries.
Anchor: residual/crash-data-loss-corruption
Fix: Hold a shared cross-process lock around the ownership read, foreign-file check, and both writes.

BLOCKING -- src/kiro_crew/mate_config.py:249 -- Recreated mates inherit their predecessor’s sidecar state

if not is_kirocrew_default_agent(source): record_initial_kit(name, source)
Trigger: Delete a mate created from a custom template after revoking an approval, then recreate the same key from kirocrew. adopt_own_file replaces the config but skips the kit record; deletion and creation leave the name-keyed revocation record intact.
Outcome: The new mate’s Built From shows the predecessor’s template, and resolve_worker_spec strips the predecessor’s revoked approvals from the replacement worker despite its fresh member identity and config.
Anchor: defect
Fix: Initialize kit and revocation state for a newly created member identity, including default-template creation; preserve existing state only for the same identity.

BLOCKING -- website/src/i18n/locales/en.json:7509 -- New roster copy reverses the recorded Mate naming decision

"memory_label_private": "Private to this crewmate",
Trigger: An English-language private-memory roster entry renders through memoryStoreLabel → KiroCrewAgentsPageRoster.
Outcome: The visible roster calls the standing teammate a “crewmate,” reversing the Mate naming decision in docs/decisions/2026-10-15-the-standing-teammate-agent-is-called-a-mate.md:8 without a superseding entry.
Anchor: recorded-decisions-are-not-regressions
Fix: Use “mate” in the new English user-facing copy; retain the existing identifiers and translation exceptions.

[BLOCK-MERGE] 1934364
[GPT-REVIEWED] 1934364

Adjudication (Opus 5.5) — is blocking on each finding proportionate?

F2 — Harm is MEDIUM. When it happens the system fails closed, quietly: a recreated mate loses approvals its owner never revoked for it, and its "Built From" label is wrong. Conditions I confirmed:

  • The kit record is keyed by name (src/kiro_crew/mate_worker_kit.py:37) and is skipped for a default-template source (src/kiro_crew/mate_config.py:249).
  • The revocations are keyed by file id (src/kiro_crew/mate_approval_revocations.py:55) and stripped in src/kiro_crew/mate_worker.py:533.
  • No delete path unlinks either record; the only references are at src/kiro_crew/mate_worker_kit.py:87,104.
  • Recovery is a manual re-grant (src/kiro_crew/mate_approval_revocations.py:67).

The real fix is small: drop both name-keyed records when a new identity is created or deleted. That costs less than the harm, so the block stands.

[ADJUDICATION] 1934364 total=1 uphold=1 downgrade=0
UPHOLD F2 src/kiro_crew/mate_config.py:249 reason=harm-warrants-remedy
[GPT-ADJUDICATED] 1934364

[ADJUDICATION-FENCED] 1934364 fenced=2 flagged=1
FLAG F1 src/kiro_crew/mate_config.py:184 -- The race needs two instances with separate data homes sharing one Kiro home, where a person creates the same mate key in both within the few milliseconds between _read_owners (mate_config.py:184) and the owner-record write (mate_config.py:192-197); once one write lands, the foreign-file check (mate_config.py:165) refuses every later attempt, so the window cannot recur by itself.
UPHOLD-FENCED F3 website/src/i18n/locales/en.json:7509 -- This PR adds the key (it is absent on the base), and it renders "crewmate" on the ordinary English private-memory roster path, against docs/decisions/2026-10-15-the-standing-teammate-agent-is-called-a-mate.md:8 with no superseding entry; that path is common, not extreme.
[GPT-ADJUDICATED-FENCED] 1934364

🏷️ Fenced finding(s) machine-flagged as likely edge case

The security fence keeps these findings blocking regardless of adjudication; the only clearance path is a human override. The annotate-only pass judged the condition combination each one requires extreme, and pre-drafted the override rationale below. A repository writer must independently verify a rationale before posting it — it is machine-authored, and a wrong override on a security-class finding ships exactly the class the fence exists to stop.

  • F1 src/kiro_crew/mate_config.py:184 — The race needs two instances with separate data homes sharing one Kiro home, where a person creates the same mate key in both within the few milliseconds between _read_owners (mate_config.py:184) and the owner-record write (mate_config.py:192-197); once one write lands, the foreign-file check (mate_config.py:165) refuses every later attempt, so the window cannot recur by itself.

    /ai-review override gpt 1934364508c03793bff44348f8c7737d271311ef: The race needs two instances with separate data homes sharing one Kiro home, where a person creates the same mate key in both within the few milliseconds between _read_owners (mate_config.py:184) and the owner-record write (mate_config.py:192-197); once one write lands, the foreign-file check (mate_config.py:165) refuses every later attempt, so the window cannot recur by itself.
    

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 1934364508c03793bff44348f8c7737d271311ef: <one-sentence reason>

@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 Oct 22, 2026
@github-actions

github-actions Bot commented Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 1934364508c03793bff44348f8c7737d271311ef — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks: the accepted RFC (#22443) settles the conductor/worker split, the one-way migration and the KAS-only worker set. The Backwards compatibility section is past the description's truncation point, so it was not judged.

Notes

  • With no switch and ACP_BACKENDS_MATE_WORKER_KIT = frozenset({ACP_BACKEND_KAS}), a crewmate whose member backend is claude, codex, opencode or goose used to hand work to workers. Now every hand-off is refused with worker_kit_harness_unsupported, and the conductor has also lost spawn. The setting is a remedy the user can flip, but the PR should name this regression and those users outright.
  • The migration is one way ("Nothing moves a crewmate back"). It repoints DM thread records to kirocrew-mate-conductor, so reverting this PR would leave threads bound to an agent that no longer exists. A one-line statement of what a revert does to migrated threads would close the rollback story.
  • The PR is very large: one commit covering about 200 files, from backend and migration through the UI and the RFC amendment. Splitting off the migration and the worker kit would make each part possible to revert on its own.

[DESIGN-REVIEWED] 1934364

Description read

[DESCRIPTION-READ] e9ac679f4cd5bc552a3b4e9b89cbad53b31ef4b8a9a1f8348a9c97c4cc54dbce

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 Oct 22, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS (a BLOCK with no Trigger and no Outcome does not block)

UX-level review of 1934364508c03793bff44348f8c7737d271311ef — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

UX-Verdict: PASS

cannot evaluate: missing the crew editor's new Advanced › "Worker's extra kit" pane and the reworked Built From section. They are the main path for giving a crewmate its own kit, and neither screenshot shows them.

Evidence gaps

  • Advanced section, WorkerKitPane. No screenshot shows the "Worker's extra kit" title and body, the "Copied from {{template}}" / "No template copied yet" line, the Template select, the "Keep what was added since the last copy" checkbox, or the "Copy into the worker" / "Re-copy from template" button. Close it with a screenshot of the crew editor open on Advanced, in both the never-copied and the already-copied state.
  • Worker-kit preview state. No screenshot shows the "Copying {{template}} would make these changes" panel, its Added/Removed/Changed rows, the Cancel/Apply buttons, the "Nothing would change." state, or the error notice. Close it with a screenshot after clicking Copy, showing the change list and Apply.
  • Built From for a crewmate that owns its config file. No screenshot shows the template-name row, the built_from_mate_hint / built_from_body_mate copy, the "Open Advanced" button, the inline error notice, or the Tools / MCP Servers / Skills counts. Close it with a screenshot of the editor on Built From.

Notes

  • The mate_upgrade_notice directions ("click my name at the top of this chat, then Edit crewmate, then Built From, in the Tools, MCP Servers, and Skills tabs") lost the blind reader. They said "I would struggle to follow them" and "I can't tell which one is 'my name at the top of this chat'". Shot-02's panel shows none of the named tabs. Replace the path with a direct link to the editor's Built From section.
  • "Worker" reads as two things. The chat says "Sent to my worker", but the worker-model help says "Workers are the extra chats Scout opens", and the reader said "I don't see where to find them". Use one definition in both places, and name where the worker sessions appear.
  • "Re-copy from template" sits beside the "Keep what was added since the last copy" checkbox. If someone unchecks the box, the copy wipes what was added since. Make the Apply label say that, e.g. "Replace the worker kit", whenever the box is unchecked.

[UX-REVIEWED] 1934364

Description read

[DESCRIPTION-READ] ba04eb372095eeff61677034a83c52d70ca57a5bde179468c91d25f774ce326a

That sha256 names everything this verdict read: the PR description, plus the 5 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 added merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Oct 22, 2026
@buluoray

Copy link
Copy Markdown
Contributor Author

Goal line updated at my request: "and git worktree" is removed. The per-crewmate git worktree was split out of this PR to keep it reviewable, and it will land in its own PR after this one. Nothing else in Goal, Why it matters or Not a goal changed.

@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 Oct 22, 2026
@buluoray

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 21c5a2d
  • What rule changed: No rule changed. Only the description text of the "agent_capabilities.py" row in _REDACTION_SINKS (src/kiro_crew/security_posture.py) changed: it now names the crewmate in-place editor (mate_capabilities) as another path through safe_view.
  • Before -> after (what was allowed/blocked, now): Before: the posture row named only the owner-facing capability rows. After: the same row also names the mate editor. What is redacted, the sink list and the masking are unchanged.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): The posture text overstates coverage, so a raw env/headers token from mates/agents/<id>.json could reach the browser without the posture page showing it. No command or login path is affected.
  • Evidence checked (which sandbox result proves it): test/test_mate_capabilities_redaction.py passes on this head. GET, preview and PUT return masked values, and a masked leaf sent back is restored from the stored file or refused, never written as the mask.
  • No regression:
    • Tools still work (which tools/commands were run, result): No tool or command path changed. Only a string literal in a tuple changed.
    • Backward compatible (old config / old data / old callers still work, how checked): The tuple shape and keys are unchanged. Readers of _REDACTION_SINKS were run (below).
    • Existing tests (which suites passed on this head): test_mate_capabilities_redaction, test_security_posture, test_mcp_panel_registration, test_session_control_parts_contract, test_session_transfer, test_transport_dispatch_composition, test_acp_frame_record, test_dashboard_agents_composition_contract: 458 passed.
  • How to undo: Revert the 3-line string change in src/kiro_crew/security_posture.py (_REDACTION_SINKS, "agent_capabilities.py" row).

@github-actions github-actions Bot removed the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 22, 2026
@buluoray

Copy link
Copy Markdown
Contributor Author

Sensitive change review

  • Head commit: 3241a33
  • What rule changed: No rule changed. Only the description text of the "agent_capabilities.py" row in _REDACTION_SINKS (src/kiro_crew/security_posture.py) changed: it now names the crewmate in-place editor (mate_capabilities) as another path through safe_view.
  • Before -> after (what was allowed/blocked, now): Before: the posture row named only the owner-facing capability rows. After: the same row also names the mate editor. What is redacted, the sink list and the masking are unchanged.
  • Worst case if wrong (breaks commands / leaks secret / locks user out): The posture text overstates coverage, so a raw env/headers token from mates/agents/<id>.json could reach the browser without the posture page showing it. No command or login path is affected.
  • Evidence checked (which sandbox result proves it): test/test_mate_capabilities_redaction.py passes on this head. GET, preview and PUT return masked values, and a masked leaf sent back is restored from the stored file or refused, never written as the mask.
  • No regression:
    • Tools still work (which tools/commands were run, result): No tool or command path changed. Only a string literal in a tuple changed.
    • Backward compatible (old config / old data / old callers still work, how checked): The tuple shape and keys are unchanged. Readers of _REDACTION_SINKS were run (below).
    • Existing tests (which suites passed on this head): test_mate_capabilities_redaction, test_security_posture, test_mcp_panel_registration, test_session_control_parts_contract, test_session_transfer, test_transport_dispatch_composition, test_acp_frame_record, test_dashboard_agents_composition_contract: 458 passed.
  • How to undo: Revert the 3-line string change in src/kiro_crew/security_posture.py (_REDACTION_SINKS, "agent_capabilities.py" row).

@buluoray

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles 3241a33: Ray Xu (crewmate design owner) decided: a crewmate needs the dispatch tools that only KAS carries among Kiro backends and kiro is moving to KAS, so a stored kiro member backend runs as KAS (logged once at load). The KAS-only worker in this PR is an interim step; the next PR, stacked on this one, runs the worker on every backend in ACP_BACKENDS_MEMBER_CAPABILITIES through the member saved-spec path.

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@buluoray marked the first-principles AI finding as false positive, not applicable, or explicitly accepted for 3241a33a91b9ef581ed17a317fc1c12a566b9533.

Ray Xu (crewmate design owner) decided: a crewmate needs the dispatch tools that only KAS carries among Kiro backends and kiro is moving to KAS, so a stored kiro member backend runs as KAS (logged once at load). The KAS-only worker in this PR is an interim step; the next PR, stacked on this one, runs the worker on every backend in ACP_BACKENDS_MEMBER_CAPABILITIES through the member saved-spec path.

This decision applies only to this commit. A new push requires a new judgment.

@iamwhatever iamwhatever mentioned this pull request Oct 24, 2026
4 of 5 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

merge conflict Branch has merge conflicts with its base — author must resolve before merge

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants