# Meshery v1.0.12

2026-04-24 · [Original release notes](https://github.com/meshery/meshery/releases/tag/v1.0.12) · [Full summary](https://ratatosk.io/en/releases/meshery/v1.0.12)

Security 1 · Breaking changes 10 · Deprecations 2 · Important fixes 5 · Enhancements 1 · Dependency updates 6 · Other changes 19

## Security (1)

### Credential ownership bound to the authenticated user

- Affects: credential save and update endpoints at `/api/integrations/credentials`

When a request included another user's UUID in `userId`, the credential could be created or updated under that user's account. The handlers now assign ownership from the authenticated user after unmarshalling, so the request body's `userId` no longer determines the owner.

> - fix(security,docs): address review feedback on #18886 @leecalcote (#18887)

## Breaking changes (10)

### Pattern and filter requests now accept canonical keys

- Affects: setups sending pattern or filter request bodies

Previously, these request bodies used snake_case field names instead of the canonical camelCase keys `patternData`, `cytoscapeJSON`, and `filterData`. Both spellings are now accepted, and the canonical key determines the parsed value when both are supplied.

Evidence: “- applies canonical-wins precedence when both spellings are present.” https://github.com/meshery/meshery/pull/18900

> - fix(remote-provider,handlers): canonical camelCase wrapper keys + dual-accept on request bodies @leecalcote (#18900)

### Workspace and environment API fields switch to camelCase

- Affects: workspace and environment resources accessed through Meshery server APIs or `mesheryctl` (Inferred)

Previously, the `v1beta1` wire fields used snake_case names such as `page_size` and `organization_id`. The migrated `v1beta3` fields use camelCase names such as `pageSize` and `organizationId`.

Evidence: “Wire form flips to canonical camelCase (`pageSize` / `totalCount` / `organizationId` / `createdAt` / `updatedAt` / `deletedAt` / `designId` / `environmentId` / `teamId` / `viewId` / `workspaceId`) on `Workspace`, `AvailableWorkspace`, `WorkspacePage`, `MesheryDesignPage`, and every mapping page.” https://github.com/meshery/meshery/pull/18889

> - chore(schemas): repoint workspace + environment consumers to canonical-casing versions (Phase 3) @leecalcote (#18889)

### Meshery resource API and CLI output schemas updated

- Affects: Component, design, and model resources accessed through the Meshery API or `mesheryctl` (Inferred)

Resource consumers move from legacy `v1beta1/*` imports to canonical-casing schema versions. API envelopes rename `Count` to `TotalCount` and accept and emit both `pageSize`/`totalCount` and legacy `page_size`/`total_count`; `mesheryctl component view` now emits `createdAt`, `updatedAt`, and `deletedAt` instead of `created_at` and `updated_at`.

Evidence: “Final Phase 3 consumer-repoint commit for meshery. Migrates the three previously-deferred resources — `component`, `pattern`/`design`, and `model` — from the legacy `v1beta1/*` import paths onto the canonical-casing versions” https://github.com/meshery/meshery/pull/18894

> - chore(schemas): repoint component + design + model consumers to canonical-casing versions (Phase 3 - final) @leecalcote (#18894)

### Connection consumers move to `v1beta3/connection` and camelCase fields

- Affects: setups using connection API data or `mesheryctl connection` command output (Inferred)

Previously, consumers used `v1beta1/connection` and fields such as `user_id`, `created_at`, `updated_at`, `deleted_at`, `page_size`, `total_count`, and `status_summary`. They now use `v1beta3/connection` and the corresponding camelCase fields: `userId`, `createdAt`, `updatedAt`, `deletedAt`, `pageSize`, `totalCount`, and `statusSummary`.

> - chore(schemas): repoint connection consumers to v1beta3 (Phase 3) @leecalcote (#18890)

### `mesheryctl connection view` omits empty fields from YAML

- Affects: YAML output from `mesheryctl connection view`

When `userId`, `deletedAt`, or `environments` is nil or empty, the YAML output no longer includes that key.

Evidence: “v1beta3 also adds `yaml:"...,omitempty"` on nullable fields on `Connection` (`userId`, `deletedAt`, `environments`). `mesheryctl connection view`'s YAML output therefore no longer emits those keys when nil/empty — the expected-output goldens are regenerated to reflect this.” https://github.com/meshery/meshery/pull/18890

> - chore(schemas): repoint connection consumers to v1beta3 (Phase 3) @leecalcote (#18890)

### Meshery resource consumers moved to `v1beta2` schemas

- Affects: Meshery or mesheryctl consumers of `schedule`, `view`, `key`, `credential`, `user`, and `organization` resources, and clients reading organization page responses (Inferred)

The organization page envelope changes `page_size` and `total_count` to `pageSize` and `totalCount`; the views-in-workspace response reader currently uses the old keys.

Evidence: “`server/models/organization.go` — import flip + canonicalized JSON tags on the Meshery-local `OrganizationsPage` envelope (`totalCount`, `pageSize`). `Page` stays `page`.” https://github.com/meshery/meshery/pull/18886

> - chore(schemas): repoint low-risk resources to canonical-casing versions (Phase 3) @leecalcote (#18886)

### Meshery wire keys changed to canonical camelCase

- Affects: integrations and UI flows using the affected wire fields or performance-profile sorting (Inferred)
- Removed in: not stated

The listed fields now use camelCase instead of snake_case, and the performance UI emits `?order=lastRun desc`. Legacy snake_case spellings remain accepted on inbound JSON during the deprecation window, with canonical values taking precedence when both are present; the implementation specifically adds this handling for the five count fields.

Evidence: “Ten wire keys move to canonical camelCase to match the schema contract; the legacy snake_case spellings remain dual-accepted on inbound JSON for the deprecation window so unmigrated producers (remote provider, Kanvas) and unmigrated query-string callers keep working. Canonical wins when both spellings are present.” https://github.com/meshery/meshery/pull/18914

> - fix: Phase 2.K cascade — flip deferred 10 wire keys (server + UI) @leecalcote (#18914)
> - fix: Phase 2.K cascade — flip deferred 10 wire keys (server + UI) @leecalcote (#18914)

### Invalid-response error code collision fixed

- Affects: mesheryctl invalid-response errors and their error-code checks

Both error codes previously used `mesheryctl-1231`, causing the MeshKit Error Codes Utility Runner CI check to fail and block PRs that required it. Invalid-response errors now use `mesheryctl-1232`, while registry publish errors keep `mesheryctl-1231`.

> - fix(mesheryctl): resolve duplicate error code mesheryctl-1231 @leecalcote (#18905)

### `mesheryctl registry publish` returns structured errors

- Affects: use of the `mesheryctl registry publish` command

When the command encountered an invalid system or a publish or cleanup error, it logged the failure and returned `nil`. It now returns an `ErrPublish` error with code `mesheryctl-1231` for those failures.

Evidence: “- err = fmt.Errorf("invalid system: %s", system) // update to meshkit + return ErrPublish(fmt.Errorf("invalid system: %s", system), system)” https://github.com/meshery/meshery/pull/18392

> - [mesheryctl] Add e2e BATS test for registry publish command @Git-HimanshuRathi (#18392)
> - [mesheryctl] Add e2e BATS test for registry publish command @Git-HimanshuRathi (#18392)

### Query parameter names changed to `orgId` and `userId`

- Affects: API callers, remote-provider requests, and UI RTK hooks using `orgID` or `user_id` query arguments

Affected handlers and remote-provider URLs previously used `orgID`, while UI RTK hooks and callers used `orgID` and/or `user_id`. For the tested workspaces handler, `?orgId=abc` succeeds while `?orgID=abc` returns 400; the UI and remote-provider requests now use the canonical camelCase names.

Evidence: “TestGetWorkspacesHandler_RequiresOrgIdCaseSensitive` — recording spy provider asserts canonical `?orgId=abc` succeeds (provider receives the value), legacy `?orgID=abc` and missing parameter both return 400.” https://github.com/meshery/meshery/pull/18875

> - fix: align identifier naming (orgID -> orgId) in handlers, remote-provider, and UI RTK hooks @leecalcote (#18875)
> - fix: align identifier naming (orgID -> orgId) in handlers, remote-provider, and UI RTK hooks @leecalcote (#18875)

## Deprecations (2)

### Pattern endpoint accepts multiple nested design field spellings

- Affects: the `POST /api/pattern` request body
- Removed in: not stated

Previously, `DesignPostPayload` accepted only `design_file` for the nested design. It now accepts `designFile`, `patternFile`, `design_file`, and `pattern_file`, with canonical spellings taking precedence; marshaling emits `designFile` and `design_file` but not the alternate `patternFile` and `pattern_file` spellings.

> - fix(handlers): dual-accept designFile/patternFile on POST /api/pattern @leecalcote (#18888)

### Request keys switch to camelCase

- Affects: Kubernetes ping, environment, event-log configuration, and workspace request flows
- Removed in: the next schemas version bump, when `v1beta2` flips the tags

The UI previously sent `connection_id`, `organization_id`, and `event_log_level` in the affected query and body fields. It now sends camelCase keys; handlers retain compatibility with legacy spellings during the deprecation window, including a fallback for the Kubernetes ping query and workspace mutation arguments.

> - fix(ui,handlers): canonical camelCase wire keys in rtk-query (Option B Phase 2 tail) @leecalcote (#18904)

## Important fixes (5)

### Imported design names preserved

- Affects: design imports (Inferred)

When an import payload had no name, the server overwrote the name produced during conversion, so the design appeared as “Autogenerated” instead of using the filename. Now it keeps the converted name, using the filename without its extension only if conversion also produced no name.

> - [Server] Use file name instead of "Autogenerated" for imported designs @guan404ming (#17415)

### [Silent failure] SaveMesheryPattern outbound wrapper key corrected

- Affects: remote-provider `SaveMesheryPattern` calls (Inferred)

On each call, meshery-cloud silently dropped the pattern body because it did not recognize `pattern_data`. The wrapper now sends `patternData`, the key expected by the cloud contract.

> - fix(remote-provider,handlers): canonical camelCase wrapper keys + dual-accept on request bodies @leecalcote (#18900)

### Connection environment results no longer clobbered

- Affects: connections with associated environments (Inferred)

When multiple returned connections had environments fetched, GORM reused the backing array for the hoisted slice, so one connection's results could overwrite another's. Each loop iteration now declares a fresh slice for its environment query.

> - chore(deps): bump meshkit to v1.0.5 + unblock master Docker build @leecalcote (#18896)

### Policy engine skips nil mutator values

- Affects: edge-non-binding, wallet, or hierarchical policies

When a mutator field resolved to `nil` in `patchMutatorsAction`, repeated identical patches could exhaust `MAX_RE_EVALUATION_DEPTH` and time out on large designs.

> - [Server] Fix policy engine infinite loop on nil mutator values @guan404ming (#18853)

### mesheryctl API response handling now reports invalid responses

- Affects: mesheryctl API operations that process HTTP responses, including model import (Inferred)

Previously, nil responses or bodies were used without a check, and body-read errors were returned directly. Nil inputs now return `ErrGenerateDataForInvalidResponse` with code `mesheryctl-1232`, while body-read failures are wrapped with `utils.ErrReadFromBody`; the helper is now named `GenerateDataFromBodyResponse`.

> - refactor(mesheryctl): update function names and improve error handling in meshery API @lekaf974 (#18631)

## Enhancements (1)

- Both key spellings emitted and accepted in `OrganizationsPage` envelopes

## Dependency updates (6)

6 more, none mentioning security

- `- chore(deps): pin schemas v1.1.1 (drop pseudo-version) @leecalcote (#18902)`
- `- chore(deps): bump meshkit to v1.0.5 + unblock master Docker build @leecalcote (#18896)`
- `- [Chore]: Update to Sistent v0.20.0 @l5io (#18912)`
- `- chore(ui): bump @sistent/sistent to v0.19.0 (Phase 2.K) @leecalcote (#18911)`
- `- chore(deps): bump @meshery/schemas to ^1.1.0 @leecalcote (#18884)`
- `- Bump uuid from 13.0.0 to 14.0.0 in /ui @[dependabot[bot]](https://github.com/apps/dependabot) (#18871)`

## Other changes (19)

- Outbound filter wrappers use `filterData`
- `EnvironmentData` alias now targets `v1beta3/environment`
- Workspace and environment validation errors return JSON
- Local user import alias renamed to `userV1beta2`
- Identifier-naming mandate and camelCase wire contract added to `AGENTS.md`
- Environment overrides for `PORT` and `USE_GO_POLICY_ENGINE` in `make server`
- Unreachable application-save methods and unused single-save helper removed
- Notification Center event-row layout corrected
- Meshery footer favorite icon color mismatch fixed
- Designs table Actions menus now open one dropdown at a time
- CodeQL scans exclude `docs/meetings/**`
- Sulagna’s introduction added to the April 23, 2026 meeting notes
- Chinmay introductory profile added
- Yash Mahakal meeting document added
- REST API documentation template resolves duplicate tag labels
- Meshery bundle OpenAPI inclusion rules documented
- REST API resources sorted alphabetically in generated documentation
- Meeting notes attendance instructions added
- CodeQL scans ignore meeting minutes under `docs/meetings/**`

---

Summaries are produced by AI, so they can be wrong, and they are provided without warranty.
