Skip to content

Access follow-ups: propose_access asks only for named roles' admins, role audit rows name the role, list_roles grants (also MCP), ?app= ghost cell, ChangeList wraps (#23) - #66

Merged
doxuta merged 10 commits into
mainfrom
slice/access-followups
Oct 27, 2026

Conversation

@doxuta

@doxuta doxuta commented Oct 27, 2026

Copy link
Copy Markdown
Owner

Closes #23

The five LOW follow-ups from #21 listed in the issue body. The items added later in the issue's comments (N-N hidden-new paths, ACL cache on table delete, PUT /admin/roles/:id without a version, VALUES() in restoreGrantSQL) are not in this PR.

Per item

  1. propose_access no longer loads every user of the tenant (edd99ca). UserRepository.PrivilegedHolders(tenant, roleIDs) asks tenant_id = ? AND role_id IN (…) AND (is_admin OR is_platform_operator) once per call, only for the roles the call names; a role that does not resolve is not asked about and is still refused in order.
    Tests: TestProposeAccess_AsksOnlyAboutTheNamedRoles (unit), TestPrivilegedHolders_OnlyAdminsAndOperatorsOfTheNamedRolesInTheTenant (real MySQL: other role, other tenant, non-privileged holder, soft-deleted admin are all left out).
  2. A role's audit rows say which role (8b18608, 6b1766e). Create, update, recreate and delete of a role write record_id = the role id; an update's diff carries the role's name on both sides, even when only grants changed (the access card) or nothing changed. 6b1766e takes the prior name from the before view instead of dereferencing priorRole, which UpdateRole still guards against nil.
    Tests: TestRole_AuditRowsNameTheRole (unit), TestRM19_AdminMetaMutations_WriteAuditRows (real MySQL, through POST/PUT /api/v1/admin/roles: the grants-only update row has the role id and name in before/after).
  3. list_roles {grants: true} lists each role's current grants, also over MCP (86c0acb, 32b4ee1, be0167f). It is a read: the tenant is the executor's, nothing is buffered, and MCP already exposes list_* tools with ReadOnlyHint. With no argument the answer is unchanged. No tool was added, and the cassette fingerprint keys on tool names only, so the demo cassettes replay unchanged. 32b4ee1 puts the description back in Vietnamese with one added sentence about grants, like the other builder tools. That keeps the MCP instructions' "Tool descriptions are in Vietnamese" true (English path for the demo screens: 421 keys measured and translated, ?lang=en, persona follows the user (#6) #58 kept the prompts Vietnamese).
    Tests: TestListRoles_WithGrants, TestListRoles_GrantsNeedTheRoleReader (unit), TestMCP_ListRolesWithGrants (real MySQL over /mcp). Tenant A's token sees only A's role and grants. Tenant B's role carries grants, including on a table with the same name, and B's token sees only B's. acl_roles, acl_role_actions and both tokens' threads are unchanged after the reads.
  4. The access page filtered with ?app= (or ?bang=) still draws the access card's ghost cell (a4afc6b, 7dc109d). accessGhost adds a column for every table the card names that the page does not draw: the real, editable column when the table is built, and the read-only "not built yet" column otherwise.
    Tests: AccessMatrix.test.tsx "a built table outside the ?app= filter gets its column…", and AccessPage.test.tsx "?app= that leaves out the card's table still draws its column and the ghost cell", which pins the page's own wiring.
  5. ChangeList lines wrap instead of being cut to one line (acf37f5). The text uses slds-hyphenate min-w-0 instead of slds-truncate, and the mark is aligned to the first line. This applies to proposal cards, the publish review and history.
    Test: ChangeList.test.tsx "a long line wraps, the mark stays on its first line". A real browser shows the computed style in the click-through below.

8380167 merges origin/main (#61, #64, #65) without conflicts.

Proof that each test catches its bug: break the fix, see red, restore

  • Item 1: asking PrivilegedHolders for every role turns TestProposeAccess_AsksOnlyAboutTheNamedRoles red (asked [[r-rep r-mgr r-ops r-own r-star]]; want one call for [r-rep r-star]). Dropping the role_id IN filter, or the tenant_id filter, from the SQL turns TestPrivilegedHolders_… red (3 holders instead of 2).
  • Item 2: the old emitMetaAudit on update turns the unit test red, and the integration test too (role update row … RecordID:<nil> … want record_id … and the role's name on both sides). The old one on create turns the integration test red (__role__ row record_id=<nil>).
  • Item 3: ignoring grants turns both unit tests red. ListRoles without its tenant pin turns TestMCP_ListRolesWithGrants red (ROLES_READ_FAILED: role "it_b_contacts_admin"). Returning no grants turns it red (tenant A's token sees [{… Grants:[]}]).
  • Item 4: accessGhost keyed on built tables (the old logic) turns both ghost tests red. AccessPage passing every entity as shown turns the AccessPage test red.
  • Item 5: slds-truncate back on the line turns the ChangeList test red.

Gates

nexus-platform$ gofmt -l .                      → (empty)
nexus-platform$ go vet ./...                    → (no output, exit 0)
nexus-platform$ go build ./cmd/server/          → exit 0
nexus-platform$ go test ./...                   → exit=0 (42 ok, no FAIL)
nexus-platform$ bash scripts/run-integration.sh -race -timeout 50m -run '(?i)(access|role|audit|mcp|propose|listroles)' ./internal/ai/ ./internal/mcpserver/ ./internal/services/ ./internal/handlers/
ok  	nexus/platform/internal/ai	4.708s
ok  	nexus/platform/internal/mcpserver	1.482s
ok  	nexus/platform/internal/services	63.252s
ok  	nexus/platform/internal/handlers	145.075s
run-integration: PASS=645 SKIP=0 go-test-exit=0
nexus-platform$ bash scripts/run-integration.sh -race -run '(?i)(privileged|role)' ./internal/auth/
ok  	nexus/platform/internal/auth	3.223s
run-integration: PASS=1 SKIP=0 go-test-exit=0
nexus-platform$ python3 scripts/check-orphan-endpoints.py --strict
✅ Không có endpoint mồ côi.
nexus-web$ npx tsc --noEmit                     → exit 0
nexus-web$ npx vitest run --changed origin/main
 Test Files  26 passed (26)
      Tests  381 passed (381)
nexus-web$ npx eslint <the 7 touched files>     → exit 0

On a fresh docker compose -p accesslow2 up --build, with the assistant replaying cassettes/demo-loop.jsonl (no model):

nexus-web$ node scripts/demo-loop.mjs <out>     (vi)
demo loop passed: 18 steps, screenshots in …/demo-loop-shots
nexus-platform$ python3 scripts/acceptance-acl.py --base http://localhost:18231/api/v1 --tenant northwind --admin admin@example.com --staff test@example.com --expect tickets
✅ ĐẠT — người thứ ba đọc được cả 1 entity yêu cầu.

verify-ui click-through on the same stack, after the loop. A draft table and an access card for it arrive over /mcp with no model. The builder previews the card, then filters the page to the helpdesk app:

PASS list_roles-over-mcp: 2 roles; crm_sales_rep on tickets = create:all, edit:owner, view:all
PASS access-card-over-mcp: quyền cho vai "crm_sales_rep": assets_33912: view(owner), create(owner), edit(owner)
PASS change-lines-wrap: 3 line(s), white-space=normal, text-overflow=clip
PASS ghost-unfiltered: column "Thiết bị 33912"
PASS ghost-under-?app=helpdesk: columns [Vai | Phiếu hỗ trợ | Thiết bị 33912], ghost on "Thiết bị 33912"
PASS rep-denied: POST /admin/mcp-tokens as the rep → 403 FORBIDDEN

ui-shots.mjs on dung/nguoi-dung-quyen and ?app=helpdesk at 1440 and 390 (admin, plus the rep, who is sent to /home): 6 shots, 0 page errors, 0 px overflow.

Not done

  • The follow-ups in the issue's comments (listed above).
  • With ?app=, the note "N bảng của app khác đang ẩn" still counts a table that the card's column now draws. It is off by one while a card is previewed.
  • The red-without-fix proofs ran in unit and integration tests. The browser click-through ran only against the fixed build.

🤖 Generated with Claude Code

doxuta and others added 10 commits October 22, 2026 12:09
…t names (#23)

ListUsers loaded every user of the tenant on each call to find whether an admin
or a platform operator holds the role. UserRepository.PrivilegedHolders asks
tenant_id = ? AND role_id IN (...) AND (is_admin OR is_platform_operator), once
per call, for the roles the call names.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…s the role (#23)

A __role__ update row had record_id NULL and, when only grants changed (the AI's
access card), no name in its diff: the audit log alone could not say which role
changed. Every role row now carries the role id as record_id, and an update's
diff has the role's name on both sides.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… also over MCP (#23)

list_roles returned names only, so the builder agent and an MCP agent could
answer "who can edit contacts?" only in part. With grants: true it returns
every role of the tenant with its grants (entity, action, access level). It is
a read: tenant pinned from the executor, nothing buffered. Without the argument
the result is unchanged, and no tool was added, so the demo cassettes replay.
The tool description is now short and in English.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…s ghost cell (#23)

With ?app= (or ?bang=) the matrix drew only that app's tables, and a built
table the previewed access card names outside it was neither drawn nor a
not-built-yet column, so its ghost cell had nowhere to go. accessGhost now adds
a column for every table the card names that the page does not draw: the
table's real, editable column when it is built, the read-only one when not.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
#23)

ChangeList truncated every line (slds-truncate), so a long rule condition, grant
or path lost the part the builder had to read, on every proposal card, the
publish review and the history. The text now wraps (slds-hyphenate, min-w-0)
with the mark on its first line. The access card keeps its sentence title: the
in-place question reads it, it was never only a workaround.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ncing a missing before-image (#23)

roleUpdateDiff took priorRole.Name while UpdateRole still guards priorRole
against nil; it now reads the name from the before view, and from the role
after the update when there is none.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ants sentence (#23)

The other builder tools and the MCP instructions ("Tool descriptions are in
Vietnamese") stay Vietnamese (#58 kept the prompts Vietnamese), so list_roles
does too; ROLES_GRANTS_UNAVAILABLE reads like its siblings. Only names are in
the cassette fingerprint, so the demo replays unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ng written (#23)

Tenant B's role now carries grants (a leak shows as a role or a grant), B's
token sees only B's, and a read leaves acl_roles, acl_role_actions and both
threads as they were.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…e and its ghost cell (#23)

AccessMatrix's test fed accessGhost directly; this one renders AccessPage with
?app=clinic and a previewed card on a table outside it, so the page's wiring
(the drawn tables as 'shown') is pinned too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@doxuta
doxuta merged commit 6e2e5b3 into main Oct 27, 2026
7 checks passed
@doxuta
doxuta deleted the slice/access-followups branch October 27, 2026 05:32
doxuta added a commit that referenced this pull request Oct 27, 2026
…n both sides (#67)

Round 1 of #71's verifier: buildDiffRows dropped every key whose value was
the same before and after. The server writes a changed sensitive (PII)
field as "[redacted]" on both sides on purpose (audit_redact.go keeps the
key: who changed what field, when), so /admin/audit stopped showing that
such a field changed: a PII-only update had no diff to open and a mixed
update lost the field's row.

Only the role writer pads an unchanged key (#66 puts the role's name on
both sides so the row names the role). buildDiffRows now drops just that
one: `name` on a __role__ row when it did not change. A rename still shows.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
doxuta added a commit that referenced this pull request Oct 27, 2026
…-only pin, dev /mcp, live agent thread, class-aware note (#22) (#68)

* fix(ai): propose_field refuses a table its tenant does not have; pin the draft-caller sibling check (#22)

propose_field left a table that is neither live nor drafted to Apply, so an
in-app turn or an MCP token could file a card for a table of another tenant
or one that does not exist. It now answers ENTITY_NOT_FOUND, as
propose_layout and propose_records do; a table proposed earlier in the turn
or the MCP thread still takes a field.

The sibling check of a dependent picklist already reads with the caller's
context since #40, so on a delivered project it judges the builder's draft.
Nothing pinned it: a unit test with a reader that shows the draft only to a
ReadDraft Caller, in a turn and through RunTool, and an integration test over
a real /mcp on a delivered project with the builder's draft value.

The option-parity cases now name a table tenant-A has; two of them proposed
onto tenant-B's table and were buffered.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(ai): pin the tool executor to read-only dependencies (#22)

MCP and the in-app builder run every tool through toolExecutor; the tool
names were pinned (read or propose_* only), but nothing stopped the executor
from gaining a dependency that writes. The pin walks what the executor is
built from, plus the probes it reaches by asserting its reader to
AccessProbeCarrier, and fails on any interface method that is not a read verb
and on any func or channel. A second test reads the package source so a new
assertion on the reader must join the walk. A third proves the pin reports a
writer.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(mcp): brake nxm_ guesses and floods; cap turns per thread and one card's size (#22)

/mcp sits outside /api/v1, so none of the API's limiters reached it: every
unauthenticated nxm_ guess cost an indexed SELECT with nothing slowing it,
and a key holder had no request budget and could grow its thread without
end, each proposal re-reading the whole thread.

middleware.MCPRateLimit sits on the /mcp route, built on the API's sliding
window. A token not yet seen working spends its client IP's lookup budget
(APIKEY_PROBE_MAX_PER_MIN, as the nxk_ brake); a token the handler answered
with a 2xx spends its own budget (DATA_RATE_LIMIT_PER_MIN, as an API key on
the data plane), so the real agents behind one address never wait on a
prober. A 401 forgets the token. Refusals are the canonical RATE_LIMITED 429.

The thread takes at most 200 turns (THREAD_FULL, reads still work) and one
proposal's row at most 256 KiB (PROPOSAL_TOO_LARGE); both refuse before
anything is written.

Tests: the brake in unit and against the real handler and token store on
MySQL; the caps in mcpserver; main.go's wiring pinned like the other brakes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(mcp): the note after a proposal follows its change class (#22)

Every MCP propose tool, and every proposal's result, told the agent that a
human "applies it to drafts and publishes it". An access card is applied at
once, behind a confirm; relationships, approvals, work queues and records
take effect when applied too.

The result note now reads the class of the card it filed
(metaversion.ClassOf on its change envelope): a draft type goes to drafts
once the project is delivered and reaches users when published; any other
takes effect as soon as it is applied. The tool descriptions and the server
instructions say only what holds for every class: a human reviews and
applies the card.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(web): next dev serves /mcp, the endpoint the Studio prints for an agent (#22)

On next dev with a relative API base, mcpEndpoint() prints the dev origin
+ /mcp, and nothing served it: a contributor copied a claude mcp add line
that could not connect. The dev proxy behind app/api/v1/[...path] moves to
lib/devProxy.ts, taking the upstream path, and app/mcp uses it too (GET,
POST, DELETE: the streamable HTTP methods). Same rules as before: Origin
stripped, the answer mirrored, 404 outside development; the demo's Caddy
already routes /mcp to the API before Next.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(studio): an open agent thread shows the cards its key adds (#22)

The builder opens an MCP key's thread from Kết nối agent (MCP); when the
agent proposed again, the open thread did not change until it was reopened.

McpConnect now hands openThread the key's expiry. Until then the thread
re-reads itself on focus and every 5 s while the tab is visible, and stops
after 10 minutes with nothing new (a focus wakes it); an expired key stops
it. Read-back turns keep their m_<seq> ids, so what the builder already
applied stays applied across a re-read. A turn sent from this tab stops the
re-read, because the server's copy would rename that turn's cards.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs(security): /mcp's brake and thread caps; #22 leaves the known gaps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(ai): PrivilegedHolders (#66) is a read; the read-only pin names it

The pin caught the method #23's fix added to the access probes. It is one
SELECT of the admins and operators holding the named roles.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(mcp): the MCP integration router carries the brake, as main.go does (#22)

Every TestMCP_ flow (scope matrix, parallel proposals, settled replay,
revocations) now runs behind the brake at its default budgets.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(mcp): a new table, field or N-N relation is hidden until publish, and the note says so (#22)

Verifier LOW on #68: the note read only ClassOf's Kind, so a
propose_relationship card (DirectUntraced, NewHidden) was told it "takes
effect as soon as it is applied", while the customer sees a new N-N relation
only after a publish, as the Studio card's footer says. A new table or field
(Draft, NewHidden) was told it "goes to drafts", but it is created hidden.

The note now has three effects: applies at once (access, approvals, work
queues, records), created at apply and hidden from the customer until a
publish (a new document of a NewHidden type), or goes to drafts (an edit of
a draft type). Pinned per envelope kind the builder tools write.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(studio): a follow read that lands after the builder sends never replaces the message (#22)

Verifier LOW on #68: the own-turn guard ran before the read. When a read was
out as the builder pressed Enter, the server copy (without the new message)
replaced the local turns and the message just sent vanished. The read now
lands through a state updater that judges the turns as they are then, and
drops itself if the builder sent a turn meanwhile.

Tests pin the race and the three guards: a hidden tab does not read and
showing it reads again, a turn of the builder's own stops the re-read, and a
late read never lands in another thread opened meanwhile. Each was seen red
with its guard removed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(ai): the read-only pin reads every type assertion and refuses GetOrCreate (#22)

Verifier LOW on #68: the source check matched only reader.(Name), so a
package-qualified or pointer target (e.reader.(services.MetadataWriter),
e.reader.(*readerWithAccessProbes).MetadataReader.(X)), an alias or a type
switch slipped past; and Get*/Find* admitted GetOrCreate and FindOrCreate.

The check now parses the package: every type assertion and type-switch case
to anything but plain data must be listed with a reason (today
AccessProbeCarrier, UserTokenLedger, *replayer). A method name passes only
if it starts with a read verb and carries no write word anywhere in it. A
sample source pins each shape the scan must catch.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(deploy): the one-command demo trusts Caddy's network, so per-IP brakes see clients (#22)

Verifier LOW on #68: behind Caddy every request reaches the API from Caddy's
address, and TRUSTED_PROXY_CIDRS was unset, so every per-IP brake (login,
SSO, failed nxk_ and nxm_ lookups) keyed one bucket for all clients: one
prober spent everyone's budget.

docker-compose.yml fixes the default network's subnet (10.213.47.0/24,
NEXUS_SUBNET to change it) and the api trusts X-Forwarded-For from it.
Caddy writes that header itself. Checked: docker compose config renders both,
and a Caddy-only stack came up on 10.213.47.0/24 with Caddy at 10.213.47.2.

Tests: behind the proxy a guesser no longer spends another client's budget,
and a wiring test keeps the two compose defaults equal and parseable.
Documented in .env.example, README and SECURITY.md.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
doxuta added a commit that referenced this pull request Oct 27, 2026
…t id pinned, audit log hides a role's unchanged name, ChangeList min-w-0 pinned (#67) (#71)

* feat(auth): RoleRepository.ListRoleGrants reads every role with its grants in one query (#67)

list_roles {grants: true} read the role list, then each role's grants
(ListActionsForRole re-reads the role first): two queries per role, and a role
deleted in between failed the whole read. ListRoleGrants is acl_roles LEFT JOIN
acl_role_actions pinned to tenant_id: one statement, one snapshot, so a role
deleted meanwhile is simply absent. Roles by name, grants by entity then
action; a role without a grant has an empty, non-nil list.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai): list_roles {grants: true} is one snapshot, bounded, and says what it cut (#67)

It now reads ListRoleGrants (one query) instead of two reads per role, so a
role deleted between the list and its grants is skipped instead of failing
with ROLES_READ_FAILED. The output is capped like snapshot_model's: at most 40
roles and 60 grants per role; a cut sets rolesTruncated/rolesTotal or
grantsTruncated/grantsTotal and a note pointing at explain_access. The issue's
tenant (20 roles x 100 tables x 4 actions) was 545 KB in one result; it is now
under 100 KB. Still a read, tenant pinned from the executor, also over MCP.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(audit): pin that RoleService.Recreate's audit row carries the role id (#67)

The undo of a role delete (publish revert) writes a create row on __role__;
its record_id is the role's id, so the audit log keeps the role's whole life
under one id. Nothing pinned it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(audit): the audit log hides a key whose value did not change, and names a role row's role (#67)

A role update's diff carries the role's name on both sides (#66) so the row
names the role; drawn as "name | X | X" it read as a rename on every grants-only
save. buildDiffRows now drops a key present on both sides with the same stored
value (compared as JSON, not as drawn: two amounts can both read "1,2 tr"), so a
save that changed nothing has no diff to open. The role's name moves to the
record column of a __role__ row, where the short id stood.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(studio): pin ChangeList's min-w-0 so a long snake_case token breaks inside the row (#67)

A token with no space cannot wrap at a space; the text is a flex item whose
minimum width is its longest word unless min-w-0 lets it shrink. Nothing pinned
the class; jsdom cannot measure, so the width is checked in a real browser too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(audit): hide only a role row's padded name, not every key equal on both sides (#67)

Round 1 of #71's verifier: buildDiffRows dropped every key whose value was
the same before and after. The server writes a changed sensitive (PII)
field as "[redacted]" on both sides on purpose (audit_redact.go keeps the
key: who changed what field, when), so /admin/audit stopped showing that
such a field changed: a PII-only update had no diff to open and a mixed
update lost the field's row.

Only the role writer pads an unchanged key (#66 puts the role's name on
both sides so the row names the role). buildDiffRows now drops just that
one: `name` on a __role__ row when it did not change. A rename still shows.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(audit): a named record cell keeps the full record id in its title (#67)

A __role__ row's record cell shows the role's name instead of the short id
(#67), and its title carried only the name, so the role id was nowhere on
screen to cross-reference. The title is now "<name> · <full record id>",
for layout rows too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(ai): pin list_roles' worst-case size and "*"-first order; the comments say what they rest on (#67)

The caps count rows, not bytes: 40 roles x 60 grants with the longest names
the schema allows measure 275 KB, not "about 160 KB"; a subtest now pins
that under 300 KB. "*" sorts before every table name only because a table
name starts with a letter (in utf8mb4_unicode_ci "_" sorts before "*");
the integration test now orders a "*" grant and an "a_first" table on real
MySQL. The deleted-meanwhile comments say the fake stands in for InnoDB's
consistent read and that a delete racing the SELECT is not exercised.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
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.

Access follow-ups (LOW)

1 participant