The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to parse and transform JSON data from external sources including local files and remote APIs. This ingestion of untrusted data creates a surface for indirect prompt injection. * Ingestion points: JSON data parsed from local files, strings, and external APIs. * Boundary markers: None; the instructions do not include specific delimiters or directives to ignore instructions embedded within the JSON content. * Capability inventory: The skill demonstrates file system access (reading and writing JSON files) and implies network access for API interactions. * Sanitization: The skill highlights structural validation using schemas (jsonschema, ajv), though this does not mitigate natural language instruction injection.
[EXTERNAL_DOWNLOADS]: The documentation suggests the use of several third-party libraries for Python and Node.js. While these are industry-standard tools for JSON processing, they represent external dependencies that would need to be installed in the environment.
[COMMAND_EXECUTION]: The skill provides numerous examples using the 'jq' command-line utility, including pipe operations (e.g., 'cat data.json | jq'), which involves shell command execution.

json-transformer