investor-call-prep

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly scrapes and ingests open/public third-party content—e.g., VC firm websites via Scrapegraph ("Step 3b: Scrape the firm's website") and web-derived results from Perplexity (Step 2 & 3c)—and uses that content to drive investor fit assessments and subsequent actions, creating a clear vector for 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 performs runtime scraping via the Scrapegraph call (orth run scrapegraph /v1/smartscraper) to fetch external sites like "https://{firm_website}" (and related paths such as /portfolio and /team) and then injects that scraped content into the agent's research/prompts, so this external content is fetched at runtime and directly controls agent outputs.