alterlab-open-notebook

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes examples and API calls that require embedding secret values verbatim (e.g., export OPEN_NOTEBOOK_ENCRYPTION_KEY and POSTing {"api_key":"sk-..."}), which would force the LLM to handle or output secrets directly.

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

• Third-party content exposure detected (high risk: 0.90). The skill's API and examples explicitly ingest open web pages via the /api/sources endpoint (SKILL.md, references/api_reference.md and examples/scripts like scripts/source_ingestion.py show adding URLs such as arXiv and Wikipedia) and the chat flow (POST /api/chat/execute with "context": {"include_sources": true}) instructs the agent to read and use that untrusted third-party content as context for LLM responses, enabling indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's Quick Start explicitly runs a runtime fetch of a remote Docker Compose file via "curl -o docker-compose.yml https://raw.githubusercontent.com/lfnovo/open-notebook/main/docker-compose.yml" and then executes it with "docker-compose up -d", meaning the fetched content directly controls which containers/code run on the host.