The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the ax CLI tool (ax organizations list and ax projects list) to discover configuration details. While this represents standard integration with the vendor's platform tools, invoking local CLI commands introduces an execution boundary that must be managed securely.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an untrusted data processing surface that could be exploited via indirect prompt injection or argument injection.
Ingestion points: Ingests user input, external logs, exported trace data, and URL contexts to retrieve names and identifiers such as {space_name_or_id}.
Boundary markers: Absent. The skill does not instruct the agent to use specific delimiters or to ignore embedded instructions within processed data or inputs.
Capability inventory: Executes local shell commands (ax) and generates markdown hyperlinks.
Sanitization: Absent. The instructions note that IDs must be base64-encoded or URL-encoded for URL generation, but they do not specify sanitization, escaping, or strict validation for variables interpolated directly into shell commands (e.g., ax projects list --space "{space_name_or_id}"), which could lead to command injection if malicious characters are input.

arize-link