The Agent Skills Directory

[SAFE]: The skill implements standard observability patterns for tracking LLM application health. It correctly references well-known services and the author's verified infrastructure without any signs of malicious intent or unauthorized data access.
[INDIRECT_PROMPT_INJECTION]: The skill defines a vulnerability surface by ingesting untrusted LLM data for monitoring purposes.
Ingestion points: LLM request messages and response content are captured through instrumentation wrappers in SKILL.md and references/implementation.md.
Boundary markers: The provided implementation code does not include explicit boundary markers or instructions to ignore embedded content within the telemetry stream.
Capability inventory: The skill environment uses openai, @langfuse/client, and @langfuse/tracing libraries and is granted Read, Write, and Edit tool permissions.
Sanitization: Captured LLM data is processed and transmitted to the observability platform without transformation or sanitization to ensure metric accuracy.

langfuse-observability