langfuse-data-handling

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests and processes trace data from an external observability platform (Langfuse), which contains untrusted inputs and outputs from LLM interactions, creating a surface for indirect prompt injection.\n
  • Ingestion points: LLM trace data and observations fetched via the @langfuse/client and langfuse API libraries.\n
  • Boundary markers: The provided code snippets do not utilize delimiters or specific instructions to the agent to disregard embedded commands in the ingested trace content.\n
  • Capability inventory: The skill possesses the capability to write data to the local file system using fs.writeFileSync and to delete records from the remote Langfuse platform.\n
  • Sanitization: The skill implements basic anonymization (SHA-256 hashing) for user and session identifiers, but it lacks sanitization mechanisms for the core trace content to prevent the execution of malicious instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 11:31 AM
Security Audit — agent-trust-hub — langfuse-data-handling