The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation instructs users to download the call.mjs script directly from the vendor's domain (https://fittings.sh/skill/call.mjs). This is a legitimate vendor resource used for handling the cryptographic handshake and API communication required for payment settlement.
[COMMAND_EXECUTION]: The skill requires the execution of the call.mjs script via node. This script interacts with the filesystem to read its own source code for an integrity check and performs network requests to settle payments.
[PROMPT_INJECTION]: The skill processes hierarchical corporate data from an external source (GLEIF via the fittings.sh proxy), presenting an indirect prompt injection surface.
Ingestion points: Data is retrieved from the https://fittings.sh/v1/lei/hierarchy endpoint within call.mjs.
Boundary markers: The script does not use explicit delimiters or instructions to isolate the retrieved data from the agent's context.
Capability inventory: The execution environment provides capabilities for network access, filesystem reading (for self-integrity verification), and signing blockchain transactions.
Sanitization: The script utilizes standard JSON.parse and JSON.stringify for data handling but does not filter the content for potentially malicious instructions embedded in the finance data.

lei-hierarchy