The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a shell script (scripts/execute.sh) and markdown instructions that utilize curl, jq, and bash to interact with IndyKite's REST APIs. This is the primary intended functionality for policy management and query execution.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to IndyKite regional API endpoints (eu.api.indykite.com and us.api.indykite.com). These are legitimate, vendor-owned domains necessary for the skill's operation.
[INDIRECT_PROMPT_INJECTION]: The scripts/execute.sh utility ingests external data to populate query parameters, creating a potential surface for indirect injection if the input files are attacker-controlled.
Ingestion points: Data is read from a user-specified JSON file or standard input in scripts/execute.sh.
Boundary markers: No specific delimiters or "ignore" instructions are used for the ingested parameter data.
Capability inventory: The script utilizes curl for network operations and cat for reading input files.
Sanitization: While the script uses printf with %q for shell-safe formatting of curl arguments, it does not perform semantic validation or sanitization of the JSON content being sent to the API.

indykite-ciq-read