The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill manages sensitive Google Ads and Google Sheets API credentials through environment variables that point to local file paths. This approach, combined with the instructions to use GitHub Secrets for storage, aligns with standard security practices for managing secrets in CI/CD environments.- [DATA_EXFILTRATION]: Network activity is strictly limited to authenticated requests directed at official Google service domains (googleads.googleapis.com and oauth2.googleapis.com) and a user-provided Slack webhook for operational reporting. No unauthorized data transmission or exfiltration vectors were observed.- [COMMAND_EXECUTION]: Budget mutations and spreadsheet interactions are performed using structured API clients and the requests library. The skill does not employ shell command execution or subprocess spawning to process data from external sources.- [REMOTE_CODE_EXECUTION]: The skill relies on well-known, versioned Python dependencies from standard registries. It does not download or execute arbitrary code from remote servers.- [PROMPT_INJECTION]: Instructions and metadata fields were analyzed for override patterns, behavior bypasses, or system prompt extraction attempts; no such patterns were detected.

shared-budget-updater