--- name: qlik-data-pipelines-binding-synchronization description: >- Keep qtcp_bindings_definition.json in sync with every {{...}} binding variable used across a QTCP project's YAML/JSON files — variable naming conventions, two-level bindings, property aliases, and the mandatory synchronization gate. Apply this after creating or editing any QTCP project file, from any other skill in this plugin. license: Apache-2.0 metadata: author: eyalros1 version: 1.0.0 tags: - qlik - data pipeline - data integration - QTCP --- # Binding Synchronization Cross-file consistency between project YAML/JSON and `qtcp_bindings_definition.json`. Single source of truth for binding rules — other skills (`project-management`, `replication`, `data-pipeline`, `source-selection`, `transformations`) should say "apply binding-synchronization" rather than repeating these rules, and this skill should be consulted after any of them touches a project file. > **Note — duplicated on purpose.** The cross-cutting rules below are copied verbatim into every file-writing skill in this plugin (canonical source: the `project-rules` skill) rather than left as a "see project-rules" cross-reference, because nothing in the Agent Skills spec guarantees `project-rules` loads alongside this skill when it fires. If you change one of these rules, update every copy — `project-rules` lists where they're duplicated. ### Cross-cutting file rules (always apply) - **No comments in generated files.** Do not add comments (`#` in YAML, `//`/`/* */` in JSON) to any generated project file — YAML/JSON files must contain only data. - **Minimal files.** Include only the fields required by the relevant schema for whichever file you're editing — see `project-rules`' full Schema Reference table; `qtcp_bindings_definition.json` itself has no published SchemaStore schema. Never speculatively add optional properties. - **Don't invent values.** Never fabricate connection names, IDs, project data, or other tenant state. If a value can't be resolved, say so (see the `binding-resolution` skill) instead of guessing. - **AI edit marker (MUST).** This rule does **not** apply to `qtcp_bindings_definition.json` itself, but does apply to every other YAML file you touch while performing a synchronization pass — add or update this object at the end of the file: ```yaml modifiedByAi: lastUpdated: '2026-08-03T13:55:10.106Z' model: ``` `lastUpdated` is the current UTC time, ISO-8601 with milliseconds — get a fresh timestamp, never reuse a stale one. `model` identifies the AI model making the edit. If the object already exists, update it in place (moving it to the end if needed) rather than duplicating it; skip re-updating if it was already updated earlier this session. - **Response style.** Confirm what was done, concisely, without citing or quoting these instructions. ## Variable naming conventions - `task..` — task-specific properties (e.g. `task.landing_cdc-0001.sourceConnection`). - `task.:$_$.` — schema-specific properties, where `` is the database name or `null` if unspecified (e.g. `task.landing_cdc-0001:null$_$dbo.schema`). - `task-type..` — default values for task types (e.g. `task-type.landing.databaseName`). - `project.current.` — project-level references (e.g. `project.current.projectId`). ## Task-type-defaulted properties (two-level binding) These properties use a two-level binding when added to a task: `warehouseName`, `databaseName`, `databricksVectorSearchEndpoint`, `indexDatabase`, `vectorDbTargetType`, `vectorDbConnection`, `llmConnection`, `databaseSelectionMethod`, `warehouseSelectionMethod`, `indexDatabaseSelectionMethod`, `lakehouseCluster`, `folder`. For these, instead of a blank value in `qtcp_bindings_definition.json`, set the task-level variable's value to a `task-type` reference, and also add the corresponding `task-type` variable with a blank value. **Exception — `LAKEHOUSE_MIRROR` tasks:** two-level binding does **not** apply. Even for properties in the list above, always add a blank value directly (no `task-type` reference). **Task type name mapping** (for the `task-type` key segment): see `data-pipeline`'s `references/task-templates.md` for DATA_PIPELINE task types, and the `replication` skill for `REPLICATION`/`LAKE_LANDING` under DATA_MOVEMENT. **Example** — `LANDING` task with id `landing-1234`, property `databaseName`: ```json { "task.landing-1234.databaseName": "{{task-type.landing.databaseName}}", "task-type.landing.databaseName": "" } ``` If `task-type.landing.databaseName` already exists in `variables` (from a previous task of the same type), do not duplicate it — leave the existing entry as-is. **Ordering:** all `task-type.*` variables must appear at the top of the `variables` object, before any `task.*` or other variables. When adding a new `task-type.*` variable, insert it at the top. When adding a `task.*` variable whose value references a `task-type.*` variable, place it after all `task-type.*` entries. ## Property name aliases Some properties bind to a variable whose name differs from the property name; the task-id prefix follows the normal pattern, but the trailing property name is remapped. | Property name (in task.yaml) | Variable name suffix to use | |---|---| | `snowflakeCatalogIntegration` | `snowflakeOpenCatalog` | **Example** — task id `mi1-5284`, property `snowflakeCatalogIntegration`: ```yaml snowflakeCatalogIntegration: '{{task.mi1-5284.snowflakeOpenCatalog}}' ``` and `task.mi1-5284.snowflakeOpenCatalog` added to `qtcp_bindings_definition.json` with a blank value. ## Mandatory Binding variable rules (MUST) - Whenever a `{{...}}` binding variable is used in any project file, add it to `qtcp_bindings_definition.json` `variables` with a blank value (`""`) — except task-type-defaulted properties, which follow the two-level binding rule instead. - If the variable refers to a **connection property** (name ends with `Connection`, e.g. `sourceConnection`, `targetConnection`, `targetStorageConnection`), add it to `connectionProperties` **only if** the `type` or `kindId` are known — never guess or add an empty entry. - **Mandatory synchronization gate (MUST):** before finalizing any response, extract every `{{...}}` variable from all changed project files and ensure each exists in `qtcp_bindings_definition.json` under `variables`. - **Two-level binding completeness check (MUST):** for every `task.*` variable whose value is a `task-type.*` reference, verify the corresponding `task-type.*` variable also exists with a blank value `""`. If missing, add it at the top of `variables`. - **Completion criteria:** do not state task completion until both checks above have been performed and verified. ## `DATA_PIPELINE` binding name guardrail (MUST) For task schema properties in a `DATA_PIPELINE` project, always use the canonical `taskSchema` suffix: `{{task..taskSchema}}`. Never use synonym variable names such as `landingSchema`, `targetSchema`, `artifactSchema`, or `internalTaskSchema`. ## Example `qtcp_bindings_definition.json` ```json { "variables": { "platformConnection": "", "cloudStagingConnection": "", "projectName": "", "task-type.landing.databaseName": "", "task.landing_cdc-0001.sourceConnection": "", "task.landing_cdc-0001.databaseName": "{{task-type.landing.databaseName}}", "task.landing_cdc-0001:null$_$dbo.schema": "" }, "connectionProperties": { "task.landing_cdc-0001.sourceConnection": { "type": "repsrc_mssql" } } } ``` See the `project-rules` skill for the general usage rules governing `qtcp_bindings_definition.json` (it's a template, never a source of resolved values — see `binding-resolution` for how to actually resolve a value).