The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions direct the agent to download a JavaScript file (call.mjs) from https://fittings.sh/skill/call.mjs and execute it using node. This pattern involves fetching and running external code at runtime.
[CREDENTIALS_UNSAFE]: The skill requires a blockchain private key to be stored in the FITTINGS_PRIVATE_KEY environment variable. The documentation acknowledges that this key is stored in plaintext, which is an unsafe practice for sensitive credentials even with the recommendation to use a throwaway wallet.
[EXTERNAL_DOWNLOADS]: The skill performs external network requests to fittings.sh to download the execution client (call.mjs) and fetch drug normalization data.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted free-text data, such as drug names from chat or scraped labels. This data is passed as the name parameter to the rxnorm.normalize tool. There are no explicit boundary markers or sanitization steps documented to prevent injection if the external API or the local execution script were to process this input unsafely.
Ingestion points: The name parameter in the GET /v1/rxnorm/normalize call accepts free-text drug names from external sources like chat and prescriptions.
Boundary markers: None provided in the instructions.
Capability inventory: The skill uses curl for metadata and node to execute a downloaded script that performs network operations and cryptographic signing.
Sanitization: No sanitization or validation of the name input is described.

fittings-rxnorm-normalize