Skip to content

fix(session-control): dispatch a worker template a crewmate shadows - #18454

Merged
bolichen97 merged 1 commit into
kirodotdev:mainfrom
rubencu:fix/session-create-agent-kind
Oct 12, 2026
Merged

bolichen97 merged 1 commit into
kirodotdev:mainfrom
rubencu:fix/session-create-agent-kind

Conversation

@rubencu

@rubencu rubencu commented Oct 9, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: A conductor can dispatch kirocrew-worker as a plain template on its own memory, so a crewmate of that name holding private memory no longer blocks every leaf dispatch with a refusal the conductor cannot act on.

Symptom (gateway 0.8.0.9, three refusals in one minute, gateway.log):

WARNING kiro_crew.dashboard.session_control: memory delegation refused for member-kirocrew-conductor: this process holds no vouched identity for the caller

The conductor, running as a crew member's DM thread, then told the user:

the gateway refused to let this DM thread create kirocrew-worker sessions because the conductor member runs on the shared default memory store while kirocrew-worker owns a private one. [...] To restore the intended path, set kirocrew-worker's memory store to default in Settings > Crew

That setting does not exist. MemoryStoreField renders a member's store read-only ("stores are never selectable here"), and PATCH /api/agents/{name} refuses a memory_store change with 409 member_memory_immutable. The user went looking for the control and found nothing.

Root cause: the pre-0.8 enrol-on-mount agent sync registered kirocrew-worker as a crew member with its own V2 store (config.agents["kirocrew-worker"].member_id set, memory_stores["member-kirocrew-worker-..."]), while kirocrew-conductor stayed on default. session_create(agent="kirocrew-worker") resolves name-first, so it selects the MEMBER. The conductor's own record is on the shared store, so the own-store admission can never agree (bind_session_execution vouches only a truthy member_id), the member DM thread is fenced, and the create is refused memory_delegation_denied — for a worker that needs no memory of its own. The refusal is terse on the wire by design (it must not confirm a guessed member name), and no shipped text named a real remedy.

Why it matters

Every goal conducted from a crew member's DM thread on such an install cannot dispatch a single leaf item through the path the charter prescribes. The conductor falls back to the default agent (as it did here) or stalls, and the user is sent to a Settings control that does not exist. Any install that ran an enrol-on-mount build before f2f5a54 carries this shape; crewmate_prune_migration deliberately never prunes kirocrew_owned rows, so the member stays.

Not a goal

What changed (motivation → approach → change)

session_create can say "the template, on my own memory", and the conductor always says it for kirocrew-worker.

Surface Change
session_create (tool, POST /api/session-control/create, create_session) optional agent_kind with one value, "template"; threaded to resolve_agent_bindings(selection_kind=...); recorded on the slot and in the birth metadata
Conductor charter (agent.py) and goal-conductor/SKILL.md always dispatch kirocrew-worker with agent_kind: "template"
memory_delegation_denied tool reply names the two real remedies and says the binding is not a setting
Server-side cause line names the shared-store caller instead of a lost vouch

The chat create route already distinguishes the two namespaces a name can live in: an explicit template pick "never pins member memory" (chat_handlers.py). session_create had no way to say the same thing, so a shadowed template was unreachable from it. dashboard/session_control.py create_session(agent_kind="") takes that one value: any other value is refused invalid_agent_kind, a kind without agent is refused agent_kind_requires_agent (an inherited caller agent is not a selection), and a stated kind that does not resolve is refused agent_unresolved naming the namespace instead of falling back to whoever answers by default. The existing template arm then does the rest: the child takes the caller's own store and the worker template's persona, as an unshadowed template child already does. The resolver's namespace verdict is recorded on the slot (agent_kind) and in the persist-at-birth metadata, so an idle template child of a shadowed name rehydrates in its namespace rather than re-resolving the bare name member-first. dashboard/handlers/session_control.py accepts the field strictly typed (invalid_field_type for a non-string); validation.py SESSION_CREATE_SCHEMA allows {template} with no default (validation runs twice per TABLE.call, and a defaulted "" would meet the allow-list on the second pass).

The charter and the skill tell the conductor to pass agent_kind: "template" on every kirocrew-worker dispatch, not to retry after a refusal: a worker needs no memory of its own, and on an install with no shadowing member the template kind resolves exactly what the bare name resolves (_resolve_agent_selection skips only the alias lookup). So no conductor meets the refusal first.

mcp_dashboard.py: agent_kind in the session_create schema and payload, with the description saying to always pass it for kirocrew-worker. A memory_delegation_denied reply appends the two real remedies (re-send with agent_kind: "template", or the owner's own tab) and says the binding is not a setting; a refused template dispatch gets the owner remedy alone; other refusals are unchanged. The folder-walk preflight refusal no longer renders a doubled Error: Error: prefix. The server-side cause line names a caller whose own record is on the shared store as such (no member id AND the default store, since a V1 crewmate on its own named store also carries no member id), before the vouched-identity causes, still naming neither store nor member.

Docs: src/kiro_crew/docs/session-control.md argument table and error rows; docs/system-specs/modules/session-control.md create-route paragraph. test_agent_refactor_contracts.py / test_agent_refactor_spec_bytes.py carry the refreshed digests for the changed conductor prompt bytes, as every prompt-changing PR does (#18111 precedent).

Backwards compatibility

Compatible: agent_kind is optional on the tool, the route and create_session; every existing call omits it and takes the unchanged name-first path. The only observable changes to an existing call are the preflight refusal text losing its doubled Error: prefix (every existing assertion is startswith("Error:") and still holds) and the appended remedy sentence on memory_delegation_denied. The chat route's agent_kind (member | template) is untouched. One intended behaviour change for the stock conductor only: on an install where the old sync enrolled kirocrew-worker as a crewmate, a conductor that could bind that crewmate (the owner's own tab) used to run leaf items on the crewmate's private memory and now runs the template on its own memory, which is what a one-item worker needs; dispatching the crewmate stays available by bare name.

Removes nothing: no field, route, code, record key or capability is removed or narrowed; the fence refuses exactly the population it refused before.

Tests

test/test_member_session_control.py::TestShadowedTemplateDispatch reproduces the host shape (a crew member's DM thread on the shared store, a worker member with a V2 store, a same-named template):

  • test_the_bare_name_is_still_refused_and_the_cause_names_the_shared_store — the fence is untouched; the log cause names the shared-store condition, not "no vouched identity", and discloses neither store nor member.
  • test_a_v1_member_on_its_own_store_is_not_logged_as_a_shared_store_caller — a memory_version-1 crewmate on its own named store (member id None, private store) naming the V2 member is refused with the vouch cause, never the shared-store cause. Mutation-proved: dropping the store check from the predicate fails it at call.
  • test_the_template_kind_dispatches_the_worker_on_the_callers_own_store — the child runs the worker template on the caller's own store, member store untouched; agent_kind recorded on the slot and in the birth metadata.
  • test_the_template_kind_rides_the_http_route — refused bare, admitted as template, invalid_field_type for a non-string kind.
  • test_a_stated_kind_is_validated_and_never_falls_back — member and crew are refused invalid_agent_kind before the resolver is consulted; agent_kind_requires_agent; agent_unresolved naming the namespace. Mutation-proved: re-admitting "member" to the closed set fails this test at call (AssertionError: member).
  • test_the_real_resolver_resolves_a_member_shadowed_name_as_the_template — the real resolve_agent_bindings (no stand-in) against a config where worker is a crew member on private store member-worker and the installed-template catalog carries a template of the same name: the bare name is refused memory_delegation_denied, the stated kind creates the child on the caller's default store with template_id == "worker" and no member_id. Mutation-proved: with selection_kind=agent_kind replaced by selection_kind="", this test fails (cannot verify delegation within the caller's memory assignment).

test/test_mcp_dashboard_folders.py::TestSessionCreateAgentKind — the kind rides the payload, is absent when omitted, and crew or member is refused by the schema before any call; the memory_delegation_denied reply carries the template and owner remedies (owner-only after a refused template dispatch), other refusals carry none; the preflight refusal carries the same remedy with no folder created.

test/test_conductor_agent.py::test_prompt_and_skill_name_the_template_dispatch_for_a_shadowed_worker — charter and skill both carry the literal "Always dispatch kirocrew-worker with agent_kind: "template"" sentence and no "Re-send".

Run locally on this head: the touched test files plus test_chat_agent_kind.py, test_chat_agent_selection.py and test_config_loader.py pass (1,141 tests); black, isort, flake8 and mypy (2,242 source files) clean.

Manual verification

Not run against a live gateway: the running install is 0.8.0.9 and this change lands in a later release. The symptom is read from the host's gateway log (three memory delegation refused for member-kirocrew-conductor: this process holds no vouched identity for the caller lines) and the host's config.json (kirocrew-conductor on default, kirocrew-worker with member_id and a member-kirocrew-worker-… V2 store, created 2026-09-15 by the enrol-on-mount sync); the unit fixture reproduces exactly that shape, and the real-resolver test runs it through the shipped resolver. The absence of the Settings control the conductor pointed at was verified in source: MemoryStoreField renders the store read-only and crew_update.py answers a memory_store change with 409 member_memory_immutable.

Related Issues

no linked issue: the refusal shape was read on a host, not filed as an issue; the two references below are adjacent fixes, not this defect.

Refs #15512 (the _created_by half of the same refusal, fixed by #15916; this PR is the shared-store half). Refs #18204 ("Not covered": a member-<slug> thread refused for a requested agent name that is itself a config.agents key — agent_kind: "template" is the way out for that shape too).

Pattern harvest

Rule candidate: review-prompt
Pattern: a deliberately terse refusal (one that withholds the store, the member, the account) must be paired with an actionable remedy at the surface the caller reads, or the caller invents one — here a Settings control that does not exist.

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 (see the Licensing section of CONTRIBUTING.md).

@rubencu
rubencu requested review from a team and KiroCrewCaptain as code owners October 9, 2026 10:50
@rubencu
rubencu requested a review from chenmingwei23 October 9, 2026 10:50
@github-actions github-actions Bot added fork Pull request from a fork (external contributor) readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Oct 9, 2026
@github-actions

github-actions Bot commented Oct 9, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5, fork) — ✅ PASS

Design-level review of 870140d712b1360b5a4ff1580418013cc1d739ae via the fork AI-review pipeline — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Check: with agent_kind: "template" always on, any non-member config.agents["kirocrew-worker"] row (model pin, workspace) on the default store now gets skipped without notice.

Suggestions

  • _resolve_agent_selection skips every config.agents hit under template, not only the private-store ones. The description lists only the private-member shadow as the behaviour change. Add the non-member row case to that note, or to the skill text, so an operator who pinned a worker's model on a crew row knows the conductor no longer reads that row.

[DESIGN-REVIEWED] 870140d

Description read

[DESCRIPTION-READ] 9821bc3d3939efcf93dc79a51bfa4bd6c672e23591001194de9cf6a3d83f8200

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 9, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5, fork) — ✅ PASS

Premise-level review of 870140d712b1360b5a4ff1580418013cc1d739ae via the fork AI-review pipeline — 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/CONCERNS are advisory.

First-Principles-Verdict: PASS

On installs where kirocrew-worker is an enrolled crewmate, owner-tab conductors now run leaf items on their own memory instead of the crewmate's. Confirm that's acceptable.

What this change ships

Inventory (8 items) — 7 justified

Intent: A conductor can dispatch the stock worker even when an older sync enrolled a crewmate with the same name. This is a FIX.

  1. session_create accepts agent_kind: "template", which runs a shadowed template on the caller's own memory — justified
  2. The conductor charter and the goal-conductor skill always dispatch kirocrew-worker as the template — justified
  3. The memory_delegation_denied tool reply names the two real remedies and says the binding is not a setting — justified
  4. The server log line names a shared-store caller instead of a "lost vouch" — justified
  5. New refusals: invalid_agent_kind, agent_kind_requires_agent, and a namespaced agent_unresolved — justified
  6. A template child's namespace survives a restart (slot and birth metadata) — justified
  7. The folder-preflight refusal loses its doubled Error: Error: prefix — rides along, harmless and declared
  8. Spec, user-doc and prompt-digest updates — justified

[FIRST-PRINCIPLES-REVIEWED] 870140d

Description read

[DESCRIPTION-READ] 681aaa18815ff1db8d1aa27802c3b21e186465160c6cbed99cbe9b104513bdf8

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 commented Oct 9, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5, fork) — ⏭️ skipped

Revision 870140d712b1360b5a4ff1580418013cc1d739ae touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Advisory — does not block merge.

@github-actions

github-actions Bot commented Oct 9, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review (fork) — ✅ no blocking findings

Reviewed 870140d712b1360b5a4ff1580418013cc1d739ae via the fork AI-review pipeline; updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 870140d

@github-actions

github-actions Bot commented Oct 9, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review (fork) — ✅ no blocking findings

Reviewed 870140d712b1360b5a4ff1580418013cc1d739ae via the fork AI-review pipeline; updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 870140d

@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 Oct 9, 2026
@rubencu
rubencu force-pushed the fix/session-create-agent-kind branch from af3e004 to 4f2509a Compare October 9, 2026 13:06
@rubencu

rubencu commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author
  • Item 2 — agent_kind: "member" has no caller, shipped for symmetry span=810a8e2f114f

fixed in 4f2509a: the closed set is template alone in create_session (agent_kind not in ("", "template")), SESSION_CREATE_SCHEMA (allowed=frozenset({"template"})) and the MCP enum (["template"]); the member-kind test is deleted.
member would only restate name-first resolution (a bare name already selects the member), so no non-test caller exists and none is added. POST /api/chat keeps its own two-value field; only the session_create surface shrank.
test_a_stated_kind_is_validated_and_never_falls_back pins the shrink: member and crew are refused invalid_agent_kind before the resolver is consulted (calls == []); re-admitting "member" fails it at call. Docs, spec and the invalid_agent_kind row say "one value, template".

@rubencu

rubencu commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author
  • Item 6 — charter retries as template only after taking the refusal span=0882198104ac

fixed in 4f2509a: the conductor charter (agent.py) and goal-conductor/SKILL.md now read "Always dispatch kirocrew-worker with agent_kind: "template""; the conditional re-send paragraph is gone from both, so no conductor meets the refusal first.
Behaviour preserving on an unshadowed install: with selection_kind="template", _resolve_agent_selection skips only the alias lookup and resolves the materialized template exactly as the bare name does; the spec paragraph records this.
test_prompt_and_skill_name_the_template_dispatch_for_a_shadowed_worker asserts the literal "Always dispatch" sentence in both texts and "Re-send" not in text. The memory_delegation_denied remedy in the tool reply stays for a caller that did send a bare name.

@rubencu

rubencu commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author
  • Suggestion — tell the conductor to always send agent_kind: "template" for kirocrew-worker span=5eecd7f4593b

fixed in 4f2509a: the charter (agent.py), goal-conductor/SKILL.md and the session_create schema description all say to always pass agent_kind: "template" for kirocrew-worker; the refuse-then-retry framing is removed, so the per-leaf refused call on affected installs no longer happens.
On an unshadowed install the template kind resolves what the bare name resolves (_resolve_agent_selection skips only the alias lookup), so the always-on form costs nothing there.

Real-resolver verification asked for in the PASS summary: test_the_real_resolver_resolves_a_member_shadowed_name_as_the_template (same head) runs the shipped resolve_agent_bindings with no stand-in against a config where worker is a crew member on private store member-worker and the installed-template catalog carries a template of the same name. The bare name is refused memory_delegation_denied; agent_kind="template" creates the child on the caller's default store with template_id == "worker" and no member_id. Replacing selection_kind=agent_kind with selection_kind="" makes that test fail.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: passed Eligible automated validation passed for the current revision readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Oct 9, 2026
@rubencu
rubencu force-pushed the fix/session-create-agent-kind branch from 4f2509a to b7d648c Compare October 9, 2026 14:03
@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 Oct 9, 2026
@rubencu

rubencu commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author
  • Shared-store cause line also matches a V1 member on its own named store span=29914e6a98b5

fixed in b7d648c: the memory_delegation_denied cause branch now requires caller_execution.member_id is None AND caller_execution.store.store_id == "default", so a memory_version-1 crewmate bound to its own named store (id None, private store, as resolve_member_execution builds it) falls through to "this process holds no vouched identity for the caller", which is true for it.
Verified the shape against resolve_member_execution (member_id = getattr(agent, "member_id", "") or None, MemoryStoreRef(store, member_id)) before fixing.
Regression test test_a_v1_member_on_its_own_store_is_not_logged_as_a_shared_store_caller builds that caller through the real config loader and resolve_member_execution, asserts the refusal, zero slots, no "shared store" in the log and no store or member name disclosed; removing the store check fails it at call. The existing shared-store cause test still passes (its caller's store_id is exactly "default").

@rubencu

rubencu commented Oct 9, 2026

Copy link
Copy Markdown
Contributor Author
  • Item 7 — folder-walk preflight refusal loses its doubled Error: prefix (rides along) span=f054c5623ac6

rebutted: not a separate cosmetic repair but a consequence of this PR's remedy rendering. _run_session_create now routes both refusal paths (the folder-walk preflight and the real create) through one _refused helper that appends _create_refusal_remedy(...) for memory_delegation_denied; the preflight path is returned to _resolve_folder_for_new_session, which prefixes Error: itself, so keeping the preflight's own prefix would render the new remedy text as Error: Error: could not create a session: ... Re-send with agent_kind ....
The change is confined to the two lines the helper replaced; every existing assertion is startswith("Error:") and still holds, and TestSessionCreateAgentKind pins the preflight refusal carrying the remedy with a single prefix.

On the summary question (an owner-tab conductor on a shadowed install used to get the enrolled kirocrew-worker crewmate with its own memory and now gets the template on the conductor's memory): intended. A one-item worker needs no memory of its own, and the charter's dispatch is the template by name; the crewmate stays reachable by a bare agent: "kirocrew-worker" for any caller that may bind it. The Backwards compatibility section of the description now states this as the one intended behaviour change.

@github-actions github-actions Bot removed the readiness: action required A blocking check or review needs attention label Oct 9, 2026
@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 11, 2026
@rubencu
rubencu force-pushed the fix/session-create-agent-kind branch from f5de819 to ec648c4 Compare October 11, 2026 16:50
@rubencu

rubencu commented Oct 11, 2026

Copy link
Copy Markdown
Contributor Author

Rebased onto main at c590eb1212 (f5de81928d -> ec648c451f), same single commit. #18854 changed the conductor system prompt too, so the two contract pins this change already updated are re-pinned from the merged bytes: _CONDUCTOR_SYSTEM_PROMPT in test_agent_refactor_contracts.py, and kirocrew-conductor.json and kirocrew-ledger-conductor.json across the eight scenarios in test_agent_refactor_spec_bytes.py. No other pin moved. This change's test files plus main's test_conductor_wake.py, test_work_ledger.py and test_chat_handlers_composition_contract.py pass: 951 tests. docs-lint, black, isort, flake8, comment-history gate and git diff --check pass.

@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: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 11, 2026
@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 12, 2026
@rubencu
rubencu force-pushed the fix/session-create-agent-kind branch from ec648c4 to 8063362 Compare October 12, 2026 01:40
@rubencu

rubencu commented Oct 12, 2026

Copy link
Copy Markdown
Contributor Author

Rebased onto main at 5639ea7c9e (ec648c451f -> 8063362a36), same single commit. #19208 changed the conductor system prompt again, so the three contract pins this change already updates are re-pinned from the merged bytes (_CONDUCTOR_SYSTEM_PROMPT, and kirocrew-conductor.json plus kirocrew-ledger-conductor.json across the eight spec-bytes scenarios); no other pin moved. This change's test files plus main's test_conductor_wake.py and test_work_ledger.py pass: 831 tests. black, isort, flake8, docs-lint, comment-history gate and git diff --check pass.

@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 12, 2026
…me-named crewmate shadows

On installs where the pre-0.8 enrol-on-mount agent sync registered
kirocrew-worker as a crew member with its own private V2 store, a bare
session_create(agent="kirocrew-worker") selects that MEMBER. A conductor
whose own record is on the shared store (a crew member's DM thread) or a
child a parent conductor created may not bind that member's store, so every
leaf dispatch is refused memory_delegation_denied -- for a worker that
needs no memory of its own. The refusal is terse by design and the member
binding is immutable, so the conductor invented a remedy that does not
exist (a memory-store setting under Settings > Crew).

session_create takes agent_kind with one value, "template": it resolves the
shared template in resolve_agent_bindings(selection_kind="template") and
the child takes the caller's own store exactly as an unshadowed template
child does. There is no "member" kind: a bare name already selects the
member. The committed namespace is recorded on the slot and in the birth
metadata. The conductor charter and the goal-conductor skill always pass
agent_kind "template" for kirocrew-worker, so no conductor meets the
refusal first; on an install with no shadowing member the template kind
resolves exactly what the bare name resolves. The MCP reply to
memory_delegation_denied names the two real remedies, the server-side
cause line names the shared-store caller instead of a lost vouch, and the
preflight refusal no longer renders a doubled Error: prefix. One test runs
the real resolver against a member-shadowed name.
@rubencu
rubencu force-pushed the fix/session-create-agent-kind branch from 8063362 to 870140d Compare October 12, 2026 03:15
@rubencu

rubencu commented Oct 12, 2026

Copy link
Copy Markdown
Contributor Author

Rebased onto main at cbf2b3614a (8063362a36 -> 870140d712), same single commit. #18758 changed the conductor prompts again, so the three contract pins this change already updates are re-pinned from the merged bytes (_CONDUCTOR_SYSTEM_PROMPT, and kirocrew-conductor.json plus kirocrew-ledger-conductor.json across the eight spec-bytes scenarios); no other pin moved. This change's test files plus main's test_conductor_wake.py and test_work_ledger.py pass: 831 tests. black, isort, flake8, docs-lint, comment-history gate and git diff --check pass.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running and removed readiness: passed Eligible automated validation passed for the current revision readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Oct 12, 2026

@bolichen97 bolichen97 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Approved at head 870140d: PR Readiness is green, no open change requests, and the review lanes passed. (AI-assisted review at bolichen97's request.)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

fork Pull request from a fork (external contributor)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants