The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands such as docker run, docker-compose, and curl for the purpose of deploying the tracing infrastructure and querying its API.
[CREDENTIALS_UNSAFE]: The docker-compose.yml template in Task E includes a hardcoded default password (zipkin_password). This is a common pattern for local development environment examples and does not constitute a leak of production secrets.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external trace data from the Zipkin service which creates a surface for indirect prompt injection.
Ingestion points: External data is ingested via curl requests to the Zipkin API in Task D.
Boundary markers: The instructions do not define boundary markers to separate service data from agent instructions.
Capability inventory: The skill has access to shell execution for docker, curl, and jq commands.
Sanitization: The skill employs jq to parse the API response, providing structural filtering of the ingested data.

zipkin