The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill accesses the sensitive environment variable KBC_STORAGE_TOKEN. This token is used to authenticate the automated POST requests to the Keboola Metastore API during the migration process.
[COMMAND_EXECUTION]: The skill executes multiple shell and Python commands to manage the migration workflow.
Detects input formats using os.listdir and os.path.isdir via inline Python.
Executes a standalone migration script (scripts/migrate.py) to process Power BI artifacts.
Performs schema validation using the jsonschema library to ensure output files meet platform requirements.
[DATA_EXFILTRATION]: The skill transmits generated semantic model data to the Keboola Metastore API (KBC_METASTORE_URL) using urllib.request. This behavior is the intended primary function of the skill to populate the vendor's repository with converted models.
[INDIRECT_PROMPT_INJECTION]: The skill processes external, user-provided Power BI artifacts (TMDL folders and JSON files) which serve as an ingestion surface for untrusted data.
Ingestion points: The skill reads files from a user-specified directory defined in INPUT_DIR (Step 0 and Step 2).
Boundary markers: There are no explicit delimiters or warnings used to prevent the agent from interpreting instructions that might be embedded within the Power BI metadata or DAX expressions.
Capability inventory: The skill possesses the ability to execute subprocesses, write files to the local directory (./out/), and perform network operations via Python's urllib module.
Sanitization: The skill mitigates risks by validating the final output against official jsonschema definitions before the data is pushed to the API.

powerbi-to-sl