The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill dynamically generates and executes SQL statements (ALTER TABLE, MERGE, INSERT) using property names and metadata fetched from the HubSpot Properties API.
Evidence: In schema.py, the generate_alter_statements function constructs SQL strings by interpolating property names. Similar patterns exist in load_bigquery.py, load_snowflake.py, and load_postgres.py for DML operations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from HubSpot API responses, creating an attack surface for indirect prompt injection.
Ingestion points: Untrusted data enters via HubSpot API responses in extract.py (records) and schema.py (property metadata).
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings for the data being processed.
Capability inventory: The skill possesses capabilities for network operations (HubSpot API), file system writes (checkpoint files in /tmp), and SQL execution against external warehouses.
Sanitization: While property names are sanitized (hyphens to underscores) before SQL interpolation, there is no explicit sanitization of property values against injection patterns.
[COMMAND_EXECUTION]: The skill constructs and executes database-specific SQL commands to synchronize data and manage schema updates.
Evidence: Implementations in load_bigquery.py, load_snowflake.py, and load_postgres.py execute generated SQL against their respective database targets.

hubspot-warehouse-sync