sapiom-agent-builder

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.95). The skill explicitly instructs obtaining and displaying a plain Sapiom API key (MCP tool returns "the plain key is shown only once") and shows examples of embedding SAPIOM_API_KEY in commands/deploy envs, which requires the agent to handle/output secret values verbatim.

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs agents to fetch and ingest open/public third-party content (e.g., scraping arbitrary web pages via Firecrawl in references/web-scraping.md and templates/llm-agent.js and querying Linkup in references/web-search.md), and then feeds that untrusted content into LLM analysis and downstream actions, so external page text could indirectly inject instructions that affect behavior.