The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's installation process fetches scripts and prompts from the author's verified GitHub repository (github.com/fourteenwm/ppc-ai-skills). These are vendor-owned resources necessary for the skill's operation.
[COMMAND_EXECUTION]: The skill executes Python scripts to perform data operations via Google Ads and Google Sheets APIs. These executions are scoped to the project directory and are invoked based on explicit user triggers for query analysis and report generation.
[PROMPT_INJECTION]: The skill ingests search terms from external advertiser accounts, creating an indirect prompt injection surface. This risk is mitigated by a mandatory 'Human Review Gate' (Step 5) where users must manually approve candidates in a spreadsheet, and a subsequent hash-based approval code requirement for any mutations.
[SAFE]: Implements a two-step mutation safety pattern for account modifications (scripts/sqr_upload_negatives.py and scripts/sqr_remove_negatives.py). This requires a deterministic approval code generated during a dry-run, preventing unauthorized or accidental automated changes to live Google Ads accounts.
[SAFE]: API credentials (google-ads.yaml, token.json) are managed as local files in the project root. The skill documentation correctly guides the user on secure storage without hardcoding secrets or exfiltrating sensitive authentication data.

sqr-pipeline