How do I install the Node Catalog Sync skill?

Run npx skillmds@latest add mims-harvard/node-catalog-sync in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Node Catalog Sync safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Node Catalog Sync?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Node Catalog Sync free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Node Catalog Sync?

mims-harvard (@mims-harvard) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Node Catalog Sync by mims-harvard — Agent Skill

Name: Node Catalog Sync
Author: mims-harvard

Node Catalog Sync

Enforce synchronization between Kedro node files and catalog YAML files in the OptimusKG project. Use when editing any Python node file under optimuskg/pipelines/*/nodes/, including modifying run() functions, changing node() inputs/outputs, adding/removing/renaming DataFrame columns, or changing column types. Also use when creating new nodes or deleting existing ones.

mims-harvard e9d8544 3.2 KB Updated Sep 17, 2026

File contents

Node–Catalog Sync

When a node file is edited, the corresponding catalog YAML files must be updated to stay in sync.

Path Mapping

Node outputs are namespace-prefixed by pipeline.py. Use this table to locate catalog files:

Layer	Node `outputs`	Catalog ID	YAML path	`filepath`
Bronze	`"{src}.{name}"`	`bronze.{src}.{name}`	`conf/base/catalog/bronze/{src}/{name}.yml`	`data/bronze/{src}/{name}.parquet`
Silver nodes	`"nodes.{entity}"`	`silver.nodes.{entity}`	`conf/base/catalog/silver/nodes/{entity}.yml`	`data/silver/nodes/{entity}.parquet`
Silver edges	`"edges.{e1}_{e2}"`	`silver.edges.{e1}_{e2}`	`conf/base/catalog/silver/edges/{e1}_{e2}.yml`	`data/silver/edges/{e1}_{e2}.parquet`
Gold	`"kg.{fmt}"`	`gold.kg.{fmt}`	`conf/base/catalog/gold/{fmt}.yml`	`data/gold/kg/{fmt}/`

Multiple outputs (list) produce one YAML file per output.

Sync Workflow

Follow all 4 steps in order after every node file edit.

Step 1: Identify affected catalog entries

Read the node() call in the edited file to find outputs (string or list).
Determine the namespace from the corresponding pipeline.py (e.g., namespace="bronze").
Construct the full catalog ID: {namespace}.{output}.
Locate the YAML file using the path mapping table above.

Step 2: Update dataset ID and filepath

Only if the outputs value in the node() call changed:

Update the YAML top-level key to the new full catalog ID.
Update filepath following the convention in the path mapping table.
Rename the YAML file to match the new output name.

Step 3: Rerun node and sync catalog

Rerun the node: uv run kedro run --nodes={node_name}
Sync schema and checksum: uv run cli sync-catalog --dataset {catalog_id}
- This reads the parquet file on disk and updates both load_args.schema and metadata.checksum in the YAML automatically.
- Use --dry-run to preview changes first.
- Use --validate to check without writing (useful in CI).
Never delete the checksum property.

Step 4: Cascade downstream

List all downstream nodes using DryRunner (no execution, just shows the DAG):

uv run kedro run --from-nodes={node_name} --runner=optimuskg.runners.DryRunner

Rerun the edited node and all its downstream dependents:
```
uv run kedro run --from-nodes={node_name}
```
Sync all affected catalog entries:
```
uv run cli sync-catalog
```

Catalog YAML Structure

{catalog_id}:
  type: optimuskg.datasets.polars.ParquetDataset
  filepath: data/{layer}/{path}.parquet
  load_args:
    schema:
      column_name: pl.Type
      struct_column:
        nested_field: pl.Type
  metadata:
    checksum: {blake2b_hex_digest}
    kedro-viz:
      layer: {layer}

---name: node-catalog-syncdescription: Enforce synchronization between Kedro node files and catalog YAML files in the OptimusKG project. Use when editing any Python node file under optimuskg/pipelines/*/nodes/, including modifying run() functions, changing node() inputs/outputs, adding/removing/renaming DataFrame columns, or changing column types. Also use when creating new nodes or deleting existing ones.---# Node–Catalog SyncWhen a node file is edited, the corresponding catalog YAML files must be updated to stay in sync.## Path MappingNode `outputs` are namespace-prefixed by `pipeline.py`. Use this table to locate catalog files:| Layer | Node `outputs` | Catalog ID | YAML path | `filepath` ||-------|---------------|------------|-----------|------------|| Bronze | `"{src}.{name}"` | `bronze.{src}.{name}` | `conf/base/catalog/bronze/{src}/{name}.yml` | `data/bronze/{src}/{name}.parquet` || Silver nodes | `"nodes.{entity}"` | `silver.nodes.{entity}` | `conf/base/catalog/silver/nodes/{entity}.yml` | `data/silver/nodes/{entity}.parquet` || Silver edges | `"edges.{e1}_{e2}"` | `silver.edges.{e1}_{e2}` | `conf/base/catalog/silver/edges/{e1}_{e2}.yml` | `data/silver/edges/{e1}_{e2}.parquet` || Gold | `"kg.{fmt}"` | `gold.kg.{fmt}` | `conf/base/catalog/gold/{fmt}.yml` | `data/gold/kg/{fmt}/` |Multiple outputs (list) produce one YAML file per output.## Sync WorkflowFollow all 4 steps in order after every node file edit.### Step 1: Identify affected catalog entries1. Read the `node()` call in the edited file to find `outputs` (string or list).2. Determine the namespace from the corresponding `pipeline.py` (e.g., `namespace="bronze"`).3. Construct the full catalog ID: `{namespace}.{output}`.4. Locate the YAML file using the path mapping table above.### Step 2: Update dataset ID and filepathOnly if the `outputs` value in the `node()` call changed:1. Update the YAML top-level key to the new full catalog ID.2. Update `filepath` following the convention in the path mapping table.3. Rename the YAML file to match the new output name.### Step 3: Rerun node and sync catalog1. Rerun the node: `uv run kedro run --nodes={node_name}`2. Sync schema and checksum: `uv run cli sync-catalog --dataset {catalog_id}` - This reads the parquet file on disk and updates both `load_args.schema` and `metadata.checksum` in the YAML automatically. - Use `--dry-run` to preview changes first. - Use `--validate` to check without writing (useful in CI).3. **Never delete the checksum property.**### Step 4: Cascade downstream1. List all downstream nodes using `DryRunner` (no execution, just shows the DAG): ``` uv run kedro run --from-nodes={node_name} --runner=optimuskg.runners.DryRunner ```2. Rerun the edited node and all its downstream dependents: ``` uv run kedro run --from-nodes={node_name} ```3. Sync all affected catalog entries: ``` uv run cli sync-catalog ```## Catalog YAML Structure```yaml{catalog_id}: type: optimuskg.datasets.polars.ParquetDataset filepath: data/{layer}/{path}.parquet load_args: schema: column_name: pl.Type struct_column: nested_field: pl.Type metadata: checksum: {blake2b_hex_digest} kedro-viz: layer: {layer}```

Frequently asked questions