ndc-package-lookup
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions prompt the agent to download a helper script (call.mjs) from the fittings.sh domain, which is an external source.
- [COMMAND_EXECUTION]: The instructions require running the downloaded JavaScript file via the node command, which constitutes execution of external code.
- [CREDENTIALS_UNSAFE]: The skill requires the user to provide a FITTINGS_PRIVATE_KEY environment variable to sign blockchain transactions. While a throwaway wallet is advised, this pattern involves placing sensitive cryptographic secrets in the agent's operating environment.
- [REMOTE_CODE_EXECUTION]: The call.mjs script includes an integrity check that fetches a hash from the vendor's server and prevents execution if the script is outdated, facilitating a manual remote update cycle.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from the fittings.sh API, which constitutes a data ingestion surface. 1. Ingestion points: Data is fetched from the vendor API in call.mjs. 2. Boundary markers: No delimiters are used to wrap the drug record content in the context. 3. Capability inventory: The skill has network access and payment signing capabilities. 4. Sanitization: No sanitization is performed on API responses beyond basic JSON parsing.
Audit Metadata