The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads its constituent scripts and documentation from the author's official GitHub repository (fourteenwm/ppc-ai-skills). These are standard vendor resources provided for the skill's functionality.
[REMOTE_CODE_EXECUTION]: Installation involves downloading Python scripts which are subsequently executed to perform audits. These scripts (portfolio_conversion_audit.py and last_conversion_dates_by_action.py) are hosted on the vendor's repository and are executed locally to interact with the Google Ads API.
[DATA_EXFILTRATION]: No unauthorized data exfiltration patterns were detected. The skill interacts with the Google Ads API using local credentials provided in google-ads.yaml. Analysis confirms that data retrieved from the API is used only for console-based reporting and is not transmitted to third-party domains.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting data from the Google Ads API.
Ingestion points: Account descriptive names and conversion action names are retrieved from the API in scripts/portfolio_conversion_audit.py and scripts/last_conversion_dates_by_action.py.
Boundary markers: None are present in the console output to separate API data from instructions.
Capability inventory: The skill uses Bash to run scripts and Read to access local files.
Sanitization: API-sourced strings are printed directly to the console without sanitization, which is a common characteristic of reporting tools.

conversion-tracking-health