Skip to content

fix(session-control): admit a chat slot bound to a member's V2 store - #11606

Merged
chenmingwei23 merged 1 commit into
mainfrom
fix/member-store-chat-slot-dispatch
Sep 18, 2026
Merged

chenmingwei23 merged 1 commit into
mainfrom
fix/member-store-chat-slot-dispatch

Conversation

@iamwhatever

@iamwhatever iamwhatever commented Sep 17, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

A crew member such as kirocrew-conductor does its work by dispatching worker
sessions: session_create, session_send, session_read_message,
session_stop, session_close. It runs not only in its pinned member-* DM
slot but also in an ordinary dashboard chat tab (chat-<n>-<ts>) bound to its
private V2 memory store. The five session-control routes admit a private member
only by the member-* key, so a conductor in a normal chat tab is refused
member_scope_denied and cannot dispatch at all — even with
agent.session_control and agent.member_dispatch both on.

Why it matters

The member operating model is unusable from the surface most people actually
open a member in. Every conductor run from a chat tab is dead on its first
session_create.

What changed (motivation → approach → change)

The gate now recognises a member by its store, not only its key. A caller
whose verified private scope is a member-owned V2 store is admitted like a
member-* DM slot, and the same creator-ownership fence bounds it: a member
reaches only the sessions it created.

Admission and the fence read the store record differently, on purpose. The
record is mutable config. An operator's own writer can un-assign the member,
drop memory_version (the loader coerces a missing key to 1), or drop the
entry — and any of those can land between the gate and authorize_target,
because the body read and the prewarms suspend in between. A fence that re-reads
the record there would fall back to bool(_created_by) and let an admitted
member reach a foreign same-workspace session. So the gate carries its
admission into the fence
: _private_caller_refusal marks the request, each
route hands the mark to session_control.py as caller_fenced, and
authorize_target takes it as precomputed_ownership_fenced=True — the
parameter close_target already uses to keep config reads out of its
synchronous re-check. The switch bypass still reads the record live at every
gate, so an un-assigned member loses its bypass at once; that direction only
tightens.

flowchart LR
  subgraph before [Before]
    G1[HTTP gate<br/>admits member]:::ctx --> B1[body read, prewarms]:::ctx --> F1[fence re-reads<br/>config record]:::removed --> R1[record flipped:<br/>unfenced]:::removed
  end
  subgraph after [After]
    G2[HTTP gate<br/>admits member]:::changed -- "carry admission" --> F2[authorize_target<br/>precomputed_ownership_fenced=True]:::added --> R2[record flipped:<br/>still creator-fenced]:::added
    G2 -.-> P2[switch bypass<br/>still reads record live]:::ctx
  end
  classDef added fill:#DCFCE7,stroke:#16A34A
  classDef changed fill:#FEF3C7,stroke:#D97706
  classDef removed fill:#FEE2E2,stroke:#DC2626,stroke-dasharray:4 3
  classDef ctx fill:#E0F2FE,stroke:#0284C7
  linkStyle 2 stroke:#DC2626,stroke-dasharray:4 3
  linkStyle 3 stroke:#16A34A,stroke-width:2px
  linkStyle 4 stroke:#16A34A,stroke-width:2px
Loading

🟩 added · 🟨 changed · 🟥 removed · 🟦 unchanged

The gate's decision travels with the request, so a config write after admission
can no longer turn a member into an unfenced caller. A refusal raised on a carried
verdict says this session can only control sessions it created itself: the
verdict says whether the caller is fenced, not which class it is, and reading the
class there would mean a config read inside close's no-suspension window.

_store_is_member_owned (dashboard/session_control.py) is one fail-closed
boolean: True only for a V2 record with a non-empty owner_member whose owner
is still an active agent bound to exactly that store. A retired or re-bound
owner, a missing record, an unreadable config or a degraded memory_stores
section all withhold member status — admission and bypass fall back under the
global switch, never wider.

A member-created worker is bound to the member's store at birth, so it is itself
a member caller and may dispatch its own children (a nested conductor). This is
intended: depth is capped by the conductor agent, not this fence, and every node
stays fenced to what it created. docs/system-specs/modules/session-control.md
records both the carried admission and this recursion.

Tests

  • test/test_session_control_member_gate.py — the gate marks a request on both
    member spellings and marks nothing for an owner or a refused caller; each of
    the four target-taking routes (stop/close/send/read) forwards the
    carried verdict to its core function (True for a member, None for an
    owner), so a route that forgets the carry fails a test; through the real
    read route with a real DashboardState, a store record that flips right
    after admission still gets 403 not_creator on a foreign tab.
  • test/test_member_session_control.py — _store_is_member_owned on every
    record shape (member, V1, ownerless, retired owner, re-bound owner, missing,
    unreadable, degraded); a flipped record (un-assign, V1 downgrade, dropped
    entry, degraded config) is refused at the fence with the carried verdict while
    the inline fence provably would not bind; a legacy V1 named-store tab keeps
    owner reach; the close re-check loads no config; a member worker dispatches a
    grandchild and is fenced off what it did not create.
  • test/test_member_memory_api.py — a member-store caller against all five
    routes, with a resolvable caller and target, refused at the exact fence code
    (memory_delegation_denied for create, not_creator for the rest), and the
    owner's tab untouched.

Manual verification

N/A — unit and route-level coverage sufficient: the route test drives the real
handler against a real DashboardState.

Related Issues

no linked issue: follow-up to the merged narrow admission (#11438); no tracking
issue was filed for the chat-tab case.

Pattern harvest

Rule candidate: review-prompt
Pattern: an authorization decision made on a verified signal at the gate is
carried into the inner check, never re-derived there from mutable state.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|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

@iamwhatever
iamwhatever requested review from a team and krishdhasmana September 17, 2026 21:23
@github-actions github-actions Bot added the readiness: action required A blocking check or review needs attention label Sep 17, 2026
@github-actions

github-actions Bot commented Sep 17, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Fable 5) — ✅ PASS

Design-level review of 76189e703f3b4e0fb42d1763564a900c42a64a8b — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Verify the operator model intends any chat tab bound to a member store — not only the pinned DM slot — to inherit the member_dispatch bypass.

Suggestions

  • The generic fence wording on a carried verdict is avoidable: the gate knows the caller class at admission time, so carry the class alongside the boolean mark and keep the member-specific refusal text without any config read in the no-suspension window.

[DESIGN-REVIEWED] 76189e7

@github-actions

github-actions Bot commented Sep 17, 2026 •

Copy link
Copy Markdown
Contributor

GPT 5.6 Review — ✅ human override accepted

Human judgment by @iamwhatever overrides the GPT 5.6 finding for 76189e703f3b4e0fb42d1763564a900c42a64a8b; the recorded reason is authoritative for this commit.

This comment is updated in place on each push.

The model was not re-run because an authorized human decision supersedes it.

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 76189e703f3b4e0fb42d1763564a900c42a64a8b: <one-sentence reason>

@github-actions

github-actions Bot commented Sep 17, 2026 •

Copy link
Copy Markdown
Contributor

Opus 4.8 Review — ✅ no blocking findings

Reviewed 76189e703f3b4e0fb42d1763564a900c42a64a8b — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 76189e7

Verdict parsed from the review's SHA-scoped output markers for commit 76189e703f3b4e0fb42d1763564a900c42a64a8b.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 76189e703f3b4e0fb42d1763564a900c42a64a8b: <one-sentence reason>

@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from 74aba1a to f31eb15 Compare September 17, 2026 21:48
@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 Sep 17, 2026
@github-actions

github-actions Bot commented Sep 17, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Fable 5) — 🟡 CONCERNS

Premise-level review of 76189e703f3b4e0fb42d1763564a900c42a64a8b — 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: CONCERNS

The switch-bypass extension to chat tabs and worker trees exceeds the reported defect — which occurred with both switches ON — and rests on DM-slot symmetry.

Not justified as shipped

  • Item 2 — inherited (symmetry): the description's defect ("refused member_scope_denied … even with agent.session_control and agent.member_dispatch both on") is fully removed by gate admission plus the carried fence; extending _member_bypass to case (b) only matters when an operator explicitly set session_control: false, and its sole support is "admitted like a member-* DM slot" — the base spec's zero-config contract names "A crew member's pinned DM slot" as its subject, and the base test pinned this exact caller refused.
    Subtraction: shrink _member_caller back to the key check (drop the state threading through _member_bypass/create_session), keeping _store_is_member_owned at the HTTP gate and the carried fence — case (b) then works at defaults and under the global switch, without piercing an operator's explicit off-switch.
    Clears when: a report or issue names a member that must dispatch from a chat tab while agent.session_control is explicitly false — the harm stated without the DM twin — or the bypass arm is shrunk as above.
  • Item 3 — rides along: nested-conductor dispatch (a member-bound worker spawning grandchildren) is declared, and fixes a sibling of the same root cause — the shipped goal-conductor skill documents nesting with depth_exceeded at the work ledger (work_ledger.py:194).

What this change ships

Inventory (8 items) — 6 justified

Intent: let a crew member running in an ordinary dashboard chat tab dispatch and control worker sessions instead of dying on its first session_create — a FIX (binding path verified at chat_handlers.py:948; the admission tests fail on base).

  1. A member agent in an ordinary chat tab can dispatch/control its workers, previously 403 member_scope_denied — justified
  2. Such a chat tab, and every worker in a member's dispatch tree, now bypasses an operator's explicit agent.session_control: false — inherited (symmetry with the DM slot; the reported defect occurred with the switch on)
  3. A member-created worker can itself dispatch grandchildren (nested conductor) — rides along (declared; same-cause sibling of the shipped skill's nested model)
  4. A config write landing mid-request can no longer unfence an admitted member (verdict carried from the gate) — justified
  5. Close's point-of-no-return re-check no longer loads config on the event loop — justified
  6. not_creator refusal text on a carried verdict is now generic ("this session can only control…") — justified
  7. A store whose owner crew was deleted or re-bound loses member status and its bypass — justified
  8. session-control.md updated in the same commit — justified

[FIRST-PRINCIPLES-REVIEWED] 76189e7

@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from 1691b01 to 82673b9 Compare September 17, 2026 22:30
@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 Sep 18, 2026
@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from 55f827f to 56d7228 Compare September 18, 2026 01:28
@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 readiness: checking Automated validation is still running labels Sep 18, 2026
@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from 56d7228 to 4588171 Compare September 18, 2026 03:02
@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 Sep 18, 2026
@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from 4588171 to a1c31db Compare September 18, 2026 03:19
@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 Sep 18, 2026
@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from a1c31db to d7669fd Compare September 18, 2026 03:47
@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 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 Sep 18, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Suggestion — mark a silently-dropped malformed memory_stores entry as degraded span=5e2860aa81c5

fixed in d2a755a by removal of the branch it targeted
_member_store_verdict and its missing-record→_STORE_UNKNOWN branch no longer exist. The fence does not classify the store record at all; it consumes the admission the HTTP gate carried (precomputed_ownership_fenced), so nothing in session control encodes what the loader does with a malformed entry.
A missing record is now simply _store_is_member_owned(...) is False, which withholds admission and the switch bypass — the fail-closed direction — with no loader change required.

@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 Sep 18, 2026
@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from d2a755a to 0ce5d27 Compare September 18, 2026 06:59
@iamwhatever

iamwhatever commented Sep 18, 2026 •

Copy link
Copy Markdown
Contributor Author

self-added: yes
mechanism: route-level test class pinning that every target-taking session-control route forwards the carried fence verdict

  • Suggestion — pin the carry convention structurally span=3978c36de8f2

fixed in 0ce5d27
TestEveryRouteForwardsTheCarriedVerdict (test_session_control_member_gate) drives each of stop / close / send / read through its real handler with the core function replaced by a recorder and asserts the recorded caller_fenced is True for an admitted member and None for an owner caller. Dropping the carry from any one route fails exactly that route's row (verified by mutation).
Chosen over a shared dispatch: the five handlers differ in what they parse before the gate (body vs query, message validation), so a single injector would add a mechanism where a test pins the same convention with none.

@iamwhatever

iamwhatever commented Sep 18, 2026 •

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Item 6 — undeclared "crew member" wording on a non-member close refusal span=738332084f49

fixed in 0ce5d27
The rider is gone rather than declared: a refusal raised on a carried verdict now reads this session can only control sessions it created itself. The verdict says whether the caller is fenced, not which class it is, and reading the class there would mean a config read inside close_target's no-suspension window, so the text names the rule. The inline path keeps its three-way wording. The PR description states this.

@iamwhatever

iamwhatever commented Sep 18, 2026 •

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Cosmetic "crew member" wording for a non-member agent-created caller on close span=29914e6a98b5

fixed in 0ce5d27
Instead of a class sentinel, the carried-verdict branch of authorize_target's fence_reason names the rule (this session can only control sessions it created itself), which is accurate for every fenced class it cannot distinguish without config: member, agent-created, and a member whose record flipped. Cron keeps its own wording (prefix-readable without config); the inline path keeps the three-way wording. not_creator and the refusal are unchanged; only the string moved.

@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 readiness: checking Automated validation is still running labels Sep 18, 2026
Let an ordinary dashboard chat slot bound to a crew member's private V2
memory store use session control, not only the member's `member-*` DM
slot. The HTTP gate admits "the caller's verified scope is a member-owned
store" and the inner ownership fence ("a member may only control sessions
it created") stays.

The gate carries the admission it made into the inner fence. The member
store record is mutable config: an operator's writer can un-assign the
member, drop `memory_version` (coerced to 1 by the loader) or drop the
entry between the gate and `authorize_target`, and a fence that re-reads
the record there would collapse to `bool(_created_by)` and let an admitted
member reach a foreign same-workspace session. `_private_caller_refusal`
marks the request on member admission; every route hands the mark to
`authorize_target` as `precomputed_ownership_fenced=True`, the same
parameter `close_target` already uses to keep config reads out of its
synchronous re-check. The switch bypass still reads the record live, so an
un-assigned member loses the bypass at once (tightening only).

`_store_is_member_owned` is one fail-closed boolean: a V2 record with a
non-empty `owner_member` whose owner is still an active agent bound to
exactly that store. A retired or re-bound owner, a missing record, an
unreadable config or a degraded `memory_stores` section all withhold
member status.

A member-created worker is itself member-bound and may dispatch its own
children (nested conductor); depth is capped by the conductor agent, not
this fence, and every node stays creator-fenced.

Updates the session-control spec and pins: the carried admission on both
member spellings; a record flip after admission (un-assign, V1 downgrade,
dropped record, degraded config) refused at the fence, unit-level and
through the real read route; the legacy V1 named-store tab keeping owner
reach; the five routes refusing a member-store caller at the exact fence
code; the close re-check reading no config; a member worker dispatching a
grandchild.
@iamwhatever
iamwhatever force-pushed the fix/member-store-chat-slot-dispatch branch from 0ce5d27 to 76189e7 Compare September 18, 2026 08:10
@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 readiness: checking Automated validation is still running labels Sep 18, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Close recheck freezes a permissive ownership verdict span=36bec42fc986

rebutted — not a defect. The close re-check deliberately freezes BOTH config-derived reads at T0: the switch via skip_enabled_check and the ownership fence via precomputed_ownership_fenced; it re-evaluates only in-memory target state (mirror / link / replacement). A mid-operation change of the CALLER's class is the same category the design already declares "not a containment boundary" for the switch.
A caller admitted UNFENCED at T0 was not a member, so it passed the switch gate only because agent.session_control was ON — under which reaching a foreign same-workspace session is authorized policy. A mid-await flip to member-owned could only ADD a fence; the frozen False grants no reach the switch had not already granted at T0.
No config-free code change exists: the caller slot's store NAME is unchanged by a record flip, so the only signal is _store_is_member_owned → KiroCrewConfig.load() — exactly the config load inside close_slot's no-suspension window that the earlier finding on this path required removing. "Abort if ownership changes before pop" reintroduces it; affirmative-only carry would over-fence V1 named-store tabs on unrelated config churn.
close_target archives to history and is reopenable; the trigger requires an operator to convert the caller's own live store into a member's V2 store inside the sub-second retirement await, gaining nothing beyond the T0 grant.
Security-fenced, so this cannot converge by rebuttal alone; maintainer reason for /ai-review override gpt 76189e703f3b4e0fb42d1763564a900c42a64a8b: — The caller-fence verdict is frozen identically across the initial admission gate and the synchronous recheck, so reusing it cannot grant reach the admission gate did not already grant under global session_control; the recheck re-reads target mirror/link/replacement live, and the close is non-destructive and reopenable.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Suggestion — carry the caller class with the mark to keep member-specific wording span=2eee4cb32097

rebutted — disproportional. Only the refusal string is at stake, and the rule-naming text is already accurate for every class a carried verdict can cover (member, agent-created, member whose record flipped).
Recovering the class means either a new class field threaded through every route and frozen through close_target's reuse — a net-new mechanism the route test would then have to pin — or a _member_caller → KiroCrewConfig.load() call inside close_slot's no-suspension window, the blocking-IO hazard skip_enabled_check and the carried verdict exist to forbid. The inline path already reads config and keeps the three-way wording.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Item 2 — case-(b) _member_bypass is an unjustified extra arm span=9917783f9b86

rebutted. The arm is not free-standing: the HTTP gate admits a member on member_dispatch OR session_control (_caller_is_member_and_reachable) and the inner gate is not session_control_enabled() and not _member_bypass(...). Dropping case (b) from _member_bypass alone recreates an admit-at-the-door / refuse-inside split — the layer-disagreement class two earlier rounds closed.
Dropping it from both layers refuses a member's chat-tab dispatch whenever an operator set session_control: false while member_dispatch (default on) stayed on — the reported defect surviving, conditioned on the switch. agent.member_dispatch is the operator's dedicated ceiling for exactly this capability; the arm keeps a member's two slot kinds symmetric under the EXISTING meaning of both switches (changing that meaning is a stated non-goal).
Maintainer note: if an explicit session_control: false should veto chat-tab members, that is a switch-semantics product call — a one-line predicate narrowing plus the matching HTTP-gate narrowing plus dropping TestMemberCreatedWorkerCanDispatch, cheaply reversible.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

self-added: yes

  • Item 3 — nested-conductor dispatch rides along span=ff1bcf6fde09

rebutted — it is the smaller surface. A member-created worker inherits the creator's agent and V2 store, so store-based recognition makes it a member caller automatically; suppressing nesting would ADD a predicate arm excluding _created_by-carrying slots.
It is declared (PR description, intent comment, spec section "A member-created worker can itself dispatch"), depth is capped by the conductor agent (goal-conductor / depth_exceeded at the work ledger), and every node stays creator-fenced; pinned by TestMemberCreatedWorkerCanDispatch.

@iamwhatever

Copy link
Copy Markdown
Contributor Author

/ai-review override gpt 76189e7: The caller-fence verdict is frozen identically across both the initial admission gate and the synchronous recheck, so reusing it in the recheck cannot grant reach the admission gate didn't already grant under global session_control; the recheck re-reads target mirror/link/replacement live, and the close is non-destructive and reopenable.

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@iamwhatever marked the gpt AI finding as false positive, not applicable, or explicitly accepted for 76189e703f3b4e0fb42d1763564a900c42a64a8b.

The caller-fence verdict is frozen identically across both the initial admission gate and the synchronous recheck, so reusing it in the recheck cannot grant reach the admission gate didn't already grant under global session_control; the recheck re-reads target mirror/link/replacement live, and the close is non-destructive and reopenable.

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

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention labels Sep 18, 2026
@chenmingwei23
chenmingwei23 merged commit b1db8e2 into main Sep 18, 2026
78 of 79 checks passed
@chenmingwei23
chenmingwei23 deleted the fix/member-store-chat-slot-dispatch branch September 18, 2026 21:32
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Sep 18, 2026
amadsalmon pushed a commit to amadsalmon/KiroCrew that referenced this pull request Sep 22, 2026
…irodotdev#11606)

Let an ordinary dashboard chat slot bound to a crew member's private V2
memory store use session control, not only the member's `member-*` DM
slot. The HTTP gate admits "the caller's verified scope is a member-owned
store" and the inner ownership fence ("a member may only control sessions
it created") stays.

The gate carries the admission it made into the inner fence. The member
store record is mutable config: an operator's writer can un-assign the
member, drop `memory_version` (coerced to 1 by the loader) or drop the
entry between the gate and `authorize_target`, and a fence that re-reads
the record there would collapse to `bool(_created_by)` and let an admitted
member reach a foreign same-workspace session. `_private_caller_refusal`
marks the request on member admission; every route hands the mark to
`authorize_target` as `precomputed_ownership_fenced=True`, the same
parameter `close_target` already uses to keep config reads out of its
synchronous re-check. The switch bypass still reads the record live, so an
un-assigned member loses the bypass at once (tightening only).

`_store_is_member_owned` is one fail-closed boolean: a V2 record with a
non-empty `owner_member` whose owner is still an active agent bound to
exactly that store. A retired or re-bound owner, a missing record, an
unreadable config or a degraded `memory_stores` section all withhold
member status.

A member-created worker is itself member-bound and may dispatch its own
children (nested conductor); depth is capped by the conductor agent, not
this fence, and every node stays creator-fenced.

Updates the session-control spec and pins: the carried admission on both
member spellings; a record flip after admission (un-assign, V1 downgrade,
dropped record, degraded config) refused at the fence, unit-level and
through the real read route; the legacy V1 named-store tab keeping owner
reach; the five routes refusing a member-store caller at the exact fence
code; the close re-check reading no config; a member worker dispatching a
grandchild.

Co-authored-by: EC2 Default User <ec2-user@ip-172-31-10-71.ec2.internal>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants