phoenix-observability
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documents the installation of official arize-phoenix packages (arize-phoenix, arize-phoenix-otel, arize-phoenix-evals, etc.) from standard registries. These are reputable tools within the AI observability ecosystem.
- [DATA_EXPOSURE]: The documentation includes examples for configuring sensitive credentials such as PHOENIX_SQL_DATABASE_URL, PHOENIX_SECRET, and SLACK_WEBHOOK_URL. It correctly advises the use of environment variables and placeholders for these secrets, adhering to security best practices for configuration management.
- [INDIRECT_PROMPT_INJECTION]: The skill inherently handles untrusted data as its primary function is to ingest and evaluate LLM traces.
- Ingestion points: External trace data is ingested via client.get_spans_dataframe() and processed in run_evals() (SKILL.md, references/advanced-usage.md).
- Boundary markers: Not explicitly defined in the provided code snippets, which are high-level examples.
- Capability inventory: The skill facilitates network communication with a tracing backend and database writes for logging; it does not demonstrate arbitrary code execution based on trace content.
- Sanitization: Standard for documentation examples, explicit sanitization logic is not shown, but the data is processed within the structured schemas of OpenInference and OpenTelemetry.
Audit Metadata