setup-email-drafts

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill prompt instructs the agent to read secrets (PSQUARED_CRM_TOKEN and EMAIL_DRAFT_ONLY_BEARER) from a .env file into shell variables and then include them verbatim in curl authorization headers (-H "Authorization: Bearer $PSQUARED_CRM_TOKEN"), which exposes the secrets to the LLM's context and output.

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

• Direct money access detected (medium risk: 0.30). The skill creates and updates email drafts, CRM tasks, and CRM opportunity statuses via API calls and GraphQL mutations (e.g. SKILL.md:369-397, 409-413, 428-434). This represents an indirect financial operation (creating outgoing outreach and billable lead/campaign activities for downstream execution). Therefore, the financial scope is indirect, which evaluates to medium severity.