pp-sumble

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs using a CLI command that embeds a token as a literal argument ("sumble-pp-cli auth set-token YOUR_TOKEN_HERE"), which requires the LLM or agent to handle and potentially output the secret verbatim (high exfiltration risk), even though an env-var alternative is mentioned.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's setup explicitly instructs running remote installers that fetch and execute code at runtime (e.g., "npx -y @mvanhorn/printing-press-library install sumble --cli-only" and "go install github.com/mvanhorn/printing-press-library/library/sales-and-crm/sumble/cmd/sumble-pp-cli@latest"), and the GitHub install URL is therefore a runtime-executed external dependency.

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

• Direct money access detected (high risk: 1.00). The skill exposes explicit credit-management commands (balance, cost-estimate, spend) and a budget-setting command that updates the account's credit ceiling ("budget set 500"). It also advertises that the CLI is "the only programmatic way to read the Sumble credit balance" and that billed commands will self-abort if the budget is exceeded. Because it provides an API to update account spend limits/credit ceilings (i.e., programmatic budget control), this is a direct financial-execution capability per the policy (analogous to "managing ad spend budgets" / updating a budget).