maple-agent-tracing-spring-ai
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes standard Java libraries from Maven Central, including
spring-boot-starter-opentelemetry,spring-ai-bom, andmicrometer-tracing-bridge-otelfor telemetry integration. - [DATA_EXFILTRATION]: The skill configures the application to export telemetry data to
https://ingest.maple.dev/v1/traces. This includes traces, metrics, and optionally the full content of agent conversations (prompts and completions) when themaple.ai.capture-contentproperty is enabled. This data transmission is the primary intended function of the tracing tool and targets the vendor's own infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill captures and exports untrusted data from agent conversations (user prompts and model outputs) to an external tracing backend.
- Ingestion points: User instructions and assistant messages are captured via
ChatModelObservationContextandChatClientObservationContextwithin theMapleAiObservationConfigclass inSKILL.md. - Boundary markers: None are implemented in the exported telemetry format by default.
- Capability inventory: The skill is limited to telemetry configuration and does not implement high-risk capabilities such as subprocess execution, direct file system modification, or arbitrary command execution.
- Sanitization: The skill provides guidance for manual redaction within the
message()helper method in the provided configuration class, although it is not enabled by default.
Audit Metadata