The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructions require the user to provide a blockchain private key via the FITTINGS_PRIVATE_KEY environment variable. The documentation acknowledges that this sensitive credential is stored in plaintext within the execution environment, posing a high risk of theft or unauthorized use.
[EXTERNAL_DOWNLOADS]: The documentation suggests that users download the primary execution script call.mjs directly from a remote URL (https://fittings.sh/skill/call.mjs) using curl, which bypasses standard package integrity checks and increases the risk of supply chain attacks.
[REMOTE_CODE_EXECUTION]: The call.mjs script performs dynamic runtime imports of several external Node.js packages, including @x402/fetch, @x402/evm, and viem, which executes remote code not bundled with the skill.
[COMMAND_EXECUTION]: The skill relies on shell command execution to run the call.mjs script and perform its core functions, creating a potential vector for command injection if inputs are not strictly controlled.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting data from external API endpoints (fittings.sh) and processing it for the agent without robust boundary markers or sanitization of the remote content.

fx-convert