airweave-setup

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt contains explicit examples and configuration that embed API keys/secrets directly into code and config files (e.g., api_key="YOUR_API_KEY", "sk_live_your_stripe_key", AIRWEAVE_API_KEY in mcp.json), which would require the agent to output secret values verbatim and creates high exfiltration risk.

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

• Third-party content exposure detected (high risk: 1.00). The skill's docs (see "Add Source Connections" in SKILL.md, the SDK-REFERENCE.md source list, and the MCP-SETUP.md MCP/tool descriptions) explicitly fetch and sync third‑party user-generated sources like Notion, Google Drive, Slack, GitHub, public webpages, etc., expose their textual_representation to searches, and support an "agentic" search/MCP loop that reads and reasons over that content—so untrusted third‑party content can be ingested and materially influence agent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs runtime fetching and execution of remote code—e.g., "git clone https://github.com/airweave-ai/airweave.git" followed by "./start.sh" and the runtime command "npx airweave-mcp-search" (which fetches and runs an npm package)—so these external sources are fetched at runtime and execute code on the host.