feat: add Claude Code agent skills for Feast

Add two agent skills following the anthropics/skills SKILL.md format: - feast-dev: Development guide for contributors covering setup, testing, linting, code style, project structure, and key abstractions - feast-feature-engineering: User-facing guide for building feature stores covering feature definitions, materialization, online/offline retrieval, on-demand transformations, and CLI reference Closes #5976 Signed-off-by: Sagar Gupta <sg85207@gmail.com>
ntkathole · Mar 16, 2026 · Mar 9, 2026 · Mar 10, 2026 · Mar 10, 2026 · Mar 10, 2026
commit 893f5f9f7405345dd7deb4ba274b648712f4b32c
diff --git a/.claude/skills/feast-dev/SKILL.md b/.claude/skills/feast-dev/SKILL.md
@@ -0,0 +1,92 @@
+---
+name: feast-dev
+description: Use this skill when contributing to the Feast codebase. Covers project setup, testing, linting, and PR workflow for the feast-dev/feast repository.
+---
+
+# Feast Development Guide
+
+## Environment Setup
+
+```bash
+# Install uv (if not installed)
+pip install uv
+
+# Create virtual environment and install Feast in editable mode with dev dependencies
+uv pip install -e ".[dev]"
+
+# Install pre-commit hooks (runs formatters and linters on commit)
+make install-precommit
+```
+
+## Running Tests
+
+### Unit Tests
+```bash
+# Run all unit tests
+make test-python-unit
+
+# Run a specific test file
+python -m pytest sdk/python/tests/unit/test_feature_store.py -v
+
+# Run a specific test
+python -m pytest sdk/python/tests/unit/test_feature_store.py::TestFeatureStore::test_apply -v
+```
+
+### Integration Tests (local)
+```bash
+# Start local test infrastructure
+make start-local-integration-tests
+
+# Run integration tests
+make test-python-integration-local
+```
+
+## Linting and Formatting
+
+```bash
+# Run all linters
+make lint
+
+# Auto-format code
+make format
+
+# Type checking
+mypy sdk/python/feast
+```
+
+## Code Style
+
+- Use type hints on all function signatures
+- Use `from __future__ import annotations` at the top of new files
+- Follow existing patterns in the module you are modifying
+- PR titles must follow semantic conventions: `feat:`, `fix:`, `ci:`, `chore:`, `docs:`
+- Add a GitHub label to PRs (e.g. `kind/bug`, `kind/feature`, `kind/housekeeping`)
+- Sign off commits with `git commit -s` (DCO requirement)
+
+## Project Structure
+
+```
+sdk/python/feast/          # Main Python SDK
+  cli.py                   # CLI entry point (feast apply, feast materialize, etc.)
+  feature_store.py         # FeatureStore class - core orchestration
+  repo_config.py           # feature_store.yaml configuration parsing
+  repo_operations.py       # feast apply / feast teardown logic
+  infra/                   # Online/offline store implementations
+    online_stores/         # Redis, DynamoDB, SQLite, etc.
+    offline_stores/        # BigQuery, Snowflake, File, etc.
+  transformation/          # On-demand and streaming transformations
+protos/feast/              # Protobuf definitions
+sdk/python/tests/          # Test suite
+  unit/                    # Fast, no external deps
+  integration/             # Requires infrastructure
+```
+
+## Key Abstractions
+
+- **FeatureStore** (`feature_store.py`): Entry point for all operations
+- **FeatureView**: Defines a set of features from a data source
+- **OnDemandFeatureView**: Computed features using request-time transformations
+- **Entity**: Join key definition (e.g. driver_id, customer_id)
+- **DataSource**: Where raw data lives (BigQuery, files, Snowflake, etc.)
+- **OnlineStore**: Low-latency feature serving (Redis, DynamoDB, SQLite)
+- **OfflineStore**: Historical feature retrieval (BigQuery, Snowflake, file)
diff --git a/.claude/skills/feast-feature-engineering/SKILL.md b/.claude/skills/feast-feature-engineering/SKILL.md
@@ -0,0 +1,149 @@
+---
+name: feast-feature-engineering
+description: Use this skill when building feature stores with Feast. Covers feature definitions, materialization, online/offline retrieval, and on-demand transformations.
+---
+
+# Feast Feature Engineering
+
+## Quick Start
+
+```bash
+pip install feast
+feast init my_project
+cd my_project/feature_repo
+feast apply
+feast ui
+```
+
+## Defining Features
+
+### feature_store.yaml
+```yaml
+project: my_project
+registry: data/registry.db
+provider: local
+online_store:
+  type: sqlite
+  path: data/online_store.db
+offline_store:
+  type: file
+entity_key_serialization_version: 3
+```
+
+### Feature Definitions (Python)
+
+```python
+from datetime import timedelta
+from feast import Entity, FeatureView, Field, FileSource
+from feast.types import Float32, Int64, String
+
+driver = Entity(
+    name="driver",
+    join_keys=["driver_id"],
+)
+
+driver_stats_source = FileSource(
+    name="driver_hourly_stats_source",
+    path="data/driver_stats.parquet",
+    timestamp_field="event_timestamp",
+)
+
+driver_stats_fv = FeatureView(
+    name="driver_hourly_stats",
+    entities=[driver],
+    schema=[
+        Field(name="conv_rate", dtype=Float32),
+        Field(name="acc_rate", dtype=Float32),
+        Field(name="avg_daily_trips", dtype=Int64),
+    ],
+    online=True,
+    source=driver_stats_source,
+    ttl=timedelta(hours=2),
+)
+```
+
+### On-Demand Feature Views (transformations)
+```python
+from feast import on_demand_feature_view, Field
+from feast.types import Float64
+import pandas as pd
+
+@on_demand_feature_view(
+    sources=[driver_stats_fv],
+    schema=[Field(name="conv_rate_plus_acc", dtype=Float64)],
+)
+def transformed_conv_rate(inputs: pd.DataFrame) -> pd.DataFrame:
+    df = pd.DataFrame()
+    df["conv_rate_plus_acc"] = inputs["conv_rate"] + inputs["acc_rate"]
+    return df
+```
+
+## Materialization
+
+```bash
+# Materialize features up to now
+feast materialize-incremental $(date -u +"%Y-%m-%dT%H:%M:%S")
+
+# Materialize a specific time range
+feast materialize 2023-01-01T00:00:00 2023-12-31T23:59:59
+```
+
+## Feature Retrieval
+
+### Historical (training data)
+```python
+from feast import FeatureStore
+import pandas as pd
+from datetime import datetime
+
+store = FeatureStore(repo_path=".")
+
+entity_df = pd.DataFrame({
+    "driver_id": [1001, 1002, 1003],
+    "event_timestamp": [datetime(2023, 5, 1)] * 3,
+})
+
+training_df = store.get_historical_features(
+    entity_df=entity_df,
+    features=[
+        "driver_hourly_stats:conv_rate",
+        "driver_hourly_stats:acc_rate",
+        "driver_hourly_stats:avg_daily_trips",
+    ],
+).to_df()
+```
+
+### Online (real-time serving)
+```python
+store = FeatureStore(repo_path=".")
+
+feature_vector = store.get_online_features(
+    features=[
+        "driver_hourly_stats:conv_rate",
+        "driver_hourly_stats:acc_rate",
+    ],
+    entity_rows=[{"driver_id": 1001}],
+).to_dict()
+```
+
+## CLI Reference
+
+| Command | Description |
+|---------|-------------|
+| `feast init [NAME]` | Create a new feature repository |
+| `feast apply` | Register feature definitions |
+| `feast teardown` | Remove all infrastructure |
+| `feast materialize START END` | Load features into online store |
+| `feast materialize-incremental END` | Incremental materialization |
+| `feast ui` | Launch web UI |
+| `feast serve` | Start online feature server |
+| `feast permissions list` | List access control rules |
+| `feast registry-dump` | Dump registry contents |
+
+## Supported Infrastructure
+
+**Online Stores**: SQLite, Redis, DynamoDB, Datastore, PostgreSQL, Cassandra, MySQL, Hazelcast, IKV
+
+**Offline Stores**: File (Parquet/Delta), BigQuery, Snowflake, Redshift, Spark, Trino, PostgreSQL, MSSQL, Clickhouse
+
+**Registries**: Local file, S3, GCS, Azure Blob, PostgreSQL, MySQL, Snowflake