The Agent Skills Directory

[DATA_EXFILTRATION]: The skill connects to the vendor's official endpoint https://mcp.observability.progress.com/mcp to retrieve observability metadata and traces. This is the intended behavior for diagnosing the Progress Observability platform setup.
[PROMPT_INJECTION]: The skill includes comprehensive instructions to mitigate indirect prompt injection from processed trace data. It explicitly directs the agent to treat trace content as data rather than instructions, defang boundary tokens (e.g., <input> to < input >), and scrub PII before echoing content.
[CREDENTIALS_UNSAFE]: The documentation references the use of the OBSERVABILITY_MCP_API_KEY environment variable and .mcp.json for authentication. This follows best practices for secret management and avoids hardcoded credentials.
[COMMAND_EXECUTION]: Provides a curl command for manual connectivity verification by the user. The command is not executed automatically by the skill and uses a shell variable for the API key.

health-check