cekura-create-agent

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs collecting provider API keys/secrets and shows examples embedding them directly in API request bodies/PATCHes (e.g., "vapi_api_key": "..."), which requires the LLM to handle and output secret values verbatim.

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly supports auto-importing production calls and auto-fetching/auto-syncing prompts and mock tools from external providers (e.g., VAPI/Retell/ElevenLabs) as documented in SKILL.md Phase 3 and Phase 4 (fields like auto_fetch_calls_enabled and auto_sync_prompt_enabled), which ingests user-generated/provider content that can directly change agent system prompts and runtime behavior.