The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill interacts with official Flexport API endpoints (api.flexport.com). It implements defensive logging practices using the pino redaction feature to ensure that sensitive information like Authorization headers and apiKey values are not written to logs. Furthermore, the instructions explicitly mandate the exclusion of shipment payloads, addresses, and other sensitive logistics documents from the telemetry stream.
[EXTERNAL_DOWNLOADS]: The skill references and utilizes standard, well-known Node.js libraries including prom-client for Prometheus metrics and pino for structured logging. These are industry-standard tools for the stated purpose of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill processes external API responses from Flexport. It mitigates injection risks by defining a strict telemetry schema that uses aggregate measurements and opaque correlation IDs, avoiding the processing of unstructured or sensitive payload data that could contain malicious instructions.
Ingestion points: External API responses from https://api.flexport.com handled in the instrumentedFlexport function.
Boundary markers: Explicit instructions to exclude shipment payloads, addresses, and credentials; requirement for aggregate outcomes only.
Capability inventory: Network requests via fetch, structured logging via pino, and metrics collection via prom-client.
Sanitization: Uses pino redaction for authorization headers and API keys.

flexport-observability