The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Node.js script (select-operations.js) to parse files. The script path is relative to the skill directory, and the input file path is provided via command-line arguments.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting data from external OpenAPI files and extracting string fields like operationId and path without sanitization. If these strings contain malicious instructions, they could attempt to influence the agent's behavior when it presents the selection options to the user.
Ingestion points: The parseOpenApiFile function in select-operations.js reads the file content from the path provided to the script.
Boundary markers: No specific delimiters or warnings are used to wrap the extracted operation data in the script's output.
Capability inventory: The skill uses node to execute a local script and fs.readFileSync to read local files.
Sanitization: The script does not perform any validation or escaping of the strings extracted from the OpenAPI document before they are returned to the agent's context.

apic-openapi-operation-selector