instagram-scraper

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow calls instagram.fetcher.sh endpoints like /api/user/search and /api/hashtag/{name}/posts, which ingest free-form user-provided query/path parameters and return Instagram content (including captions/comments) to the LLM.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill exposes and points agents at a live MCP endpoint (https://instagram.fetcher.sh/mcp) which an MCP-connected agent can call at runtime to discover tools and fetch_data, allowing remote content to directly influence agent behavior.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill exposes explicit payment-related operations: it references a POST /api/credits/topup, lists a paid MCP method topup_credits, and supports x402 pay-per-call using USDC on specific blockchains (Base, Polygon, Arbitrum, Monad, Solana) which implies on-chain signing/payment. These are specific crypto/payment integrations (not generic HTTP/click tooling), so the skill grants direct financial execution capability.