The Agent Skills Directory

[SAFE]: The skill follows secure practices by advising the use of environment variables for API keys and database credentials. Placeholders are used in documentation examples (e.g., in the Docker Compose section).
[COMMAND_EXECUTION]: Employs curl and jq for API interaction and JSON processing. This is consistent with the skill's purpose of querying a REST API.
[DATA_EXFILTRATION]: Interacts with a user-defined Langfuse host to retrieve observability data. No patterns of unauthorized exfiltration or exposure of local sensitive files were found.
[SAFE]: The skill ingests data (trace inputs and outputs) from the Langfuse API which may contain untrusted content from previous LLM interactions. While this creates a potential surface for indirect prompt injection, it is inherent to the function of an observability tool and no exploitable capabilities were found.
Ingestion points: API responses from endpoints like /api/public/traces/TRACE_ID (SKILL.md).
Boundary markers: None present; content is parsed by jq for display.
Capability inventory: Execution of curl and jq commands for data retrieval.
Sanitization: Relies on jq for structural parsing of retrieved JSON data.

langfuse-observability