setup-tools

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes explicit API tokens in the example tool configs (e.g., "Bearer sk_live_abc123", "Bearer helpdesk_key_789") and instructs placing an api_token value in the tool's request, which forces the LLM to handle or emit secret values verbatim in generated tool/config payloads—creating an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md and reference docs (e.g., references/world-time.md and the custom-tool examples in setup-tools and references/tool-schemas.md) explicitly instruct the agent to fetch and ingest responses from third‑party HTTP endpoints (WorldTimeAPI, Cal.com, arbitrary URLs via custom_task / generated-from-cURL), and Bolna feeds those API responses back into the conversation for the LLM to read and act on, so external/untrusted content can materially influence tool use and next actions.