build-agent

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly tells the agent to ask the user for their AGENTA_API_KEY and either create a .env containing AGENTA_API_KEY=... or "hand them a block to paste" (export/.env contents), which requires the LLM to receive and emit the secret value verbatim.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). High likelihood: the required runtime path for event-driven agents is create-subscription.sh → platform delivers provider event payload into the run inputs (references/trigger-inputs.md), and that payload is outsider-authored free text (e.g., GitHub issue body / Slack message) which the agent LLM then consumes as inputs.messages/event.attributes.