The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's documentation instructs users to download and execute a helper script (call.mjs) from the vendor's domain (https://fittings.sh). This script is required to handle the x402 payment protocol for API access.
[CREDENTIALS_UNSAFE]: The skill requires a blockchain private key to be set as the FITTINGS_PRIVATE_KEY environment variable. The author includes explicit warnings for users to use low-value throwaway wallets and notes that the key is stored in plaintext where the agent runs.
[EXTERNAL_DOWNLOADS]: The skill fetches its primary execution script and an integrity hash from the vendor's server to ensure the client-side signature logic is synchronized with the gateway.
[COMMAND_EXECUTION]: The documentation provides specific shell commands for the agent or user to execute for both dry runs and live drug lookup tasks.
[PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection by processing external drug identifiers and API responses.
Ingestion points: The drug and application parameters ingested via SKILL.md are used to build the API request.
Boundary markers: There are no explicit markers or instructions telling the agent to ignore directives within the returned drug data.
Capability inventory: The skill has network access via fetch and the ability to sign transactions.
Sanitization: While the script validates argument types (e.g., numbers, booleans), it does not perform content sanitization on the external data retrieved from the FDA data endpoint before presenting it to the agent.

orangebook-therapeutic-equivalence