speckle-core-api

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflow (SKILL.md and references/examples.md) instructs the agent to query arbitrary Speckle GraphQL endpoints (e.g., serverInfo, project/model/version) and to download object trees via REST (e.g., https://app.speckle.systems/objects//), meaning it ingests untrusted, user-generated content from public or self-hosted servers that could materially influence actions.