crossref-search

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). These URLs point to a single domain that serves an executable JavaScript module which the README explicitly tells users to curl and run (curl -sO https://fittings.sh/skill/call.mjs + node call.mjs), so fetching and executing code directly from this untrusted host is a potential malware distribution vector.

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path call.mjs builds a request from user-supplied query (argv) to GET https://fittings.sh/v1/crossref/search, and the client then ingests the returned JSON/text from that endpoint.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs fetching and then running remote executable code via "curl -sO https://fittings.sh/skill/call.mjs" (a required client script if missing), so this URL supplies code that would be executed by the operator at runtime.

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

• Direct money access detected (high risk: 1.00). The skill explicitly requires and documents on-chain payment signing: it returns a 402 with a signed quote that the client must sign, shows an example exporting FITTINGS_PRIVATE_KEY, references USDC on Base, and points to @x402/evm tooling. Those are concrete crypto wallet/signing instructions enabling the agent to authorize/pay with a private key (direct financial execution).