tool-openclaw

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

• Third-party content exposure detected (high risk: 0.90). The skill’s docs and workflow explicitly accept and ingest external, user-generated content — e.g., the webhook agent endpoint and mappings (references/docs/automation/webhook.md) and the Gmail Pub/Sub integration that templates incoming emails into agent messages (references/docs/automation/gmail-pubsub.md) — so untrusted third-party payloads can be fed into and drive the agent’s prompts and actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill supports refreshing its bundled docs snapshot by fetching https://docs.openclaw.ai/llms.txt (and the linked https://docs.openclaw.ai/*.md pages) at runtime, and that fetched content becomes the agent's source-of-truth used to craft prompts/responses, so remote content fetched from that URL can directly control agent instructions.