The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes trace data that may contain untrusted content from third-party LLM applications.
Ingestion points: Trace and span data are imported via ax spans export and ax traces export commands as described in SKILL.md.
Boundary markers: The skill includes a specific 'untrusted content guardrail' warning instructing the agent to never execute or act on content found within span attributes.
Capability inventory: The skill allows shell command execution (ax, jq, date) and network operations via the Arize CLI.
Sanitization: The skill relies on natural language instructions to ensure the agent treats external data as raw text for analysis only.
[COMMAND_EXECUTION]: The skill requires the use of the ax CLI tool and standard system utilities like jq and date to manage and inspect trace data.
[EXTERNAL_DOWNLOADS]: The skill depends on the arize-ax-cli package, which is the official command-line interface provided by the vendor (Arize AI). Instructions for installation via PyPI/uv are provided in references/ax-setup.md.

arize-trace