google-ads-api-setup

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly prints and instructs the user to copy/paste refresh tokens and to fill secret fields (client_secret, refresh_token, developer_token) into a YAML file, which requires handling secret values verbatim and therefore presents an exfiltration risk.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill's Installation step clones a remote Git repository (https://github.com/fourteenwm/ppc-ai-skills.git) which fetches scripts that the user is instructed to run locally (e.g., python generate_credentials.py), meaning remote code is retrieved and executed as part of setup.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). This skill specifically provisions Google Ads API credentials (generates refresh tokens, produces google-ads.yaml, and verifies connectivity) that grant read/write access to linked Google Ads accounts. Google Ads API access can be used to modify ad accounts and budgets (i.e., change ad spend), and the README explicitly states this is a prerequisite for skills that "query or modify Google Ads accounts" and warns that refresh tokens "grant full access to linked Google Ads accounts." Although the skill itself is a setup step (not an immediate money-transfer call), it explicitly enables programmatic management of ad accounts/budgets — a direct financial execution capability per the policy (managing ad spend budgets).