Security audit
Security checks for vulnerabilities and agentic risk
This Langfuse skill is a documentation-only integration helper whose network, credential, tracing, and mutation guidance is disclosed and aligned with managing a Langfuse project.
Installers should understand that this skill helps an agent send prompts, traces, scores, and related project data to Langfuse or a self-hosted Langfuse endpoint. Use scoped project keys, prefer environment variables over pasted secrets, avoid direct personal identifiers in trace metadata when possible, and review generated mutations such as prompt creation or score writes before running them against production projects.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
curl -X POST "https://cloud.langfuse.com/api/public/v2/prompts" \
-u "your-public-key:your-secret-key" \
-H "Content-Type: application/json" \
-d '{
The guidance explicitly recommends attaching user ID and session ID to tracing metadata but does not include any minimization, consent, retention, or sensitivity warning. In an observability/tracing skill, this can normalize logging identifiers into a third-party telemetry system and increase privacy, compliance, and correlation risk if implementers send direct identifiers or mix in additional personal data.
The example shows Basic Auth with curl -u public-key:secret-key ..., which encourages placing credentials directly on the command line. Command-line secrets may be exposed via shell history, process listings, logs, screenshots, or copied documentation, making accidental credential disclosure more likely.
No suspicious patterns detected.