lastXdays

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests public, user-generated content from Reddit and X (via openai_reddit and xai_x in SPEC.md and scripts/lastXdays.py) and also directs Claude to run WebSearch and synthesize those web results with Reddit/X (see the "### WEBSEARCH REQUIRED ###" instructions in scripts/lastXdays.py and the websearch plan), so untrusted third‑party content is read and can influence ranking, synthesis, and generated actions.

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

• Potentially malicious external URL detected (high risk: 0.80). This skill explicitly invokes Claude's WebSearch tool at runtime (see https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool) to fetch arbitrary external web pages whose contents are then synthesized into the agent's context (used to control outputs), and the skill can operate in a web-only mode that relies on those fetched pages.