The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx langwatch report to fetch and run the vendor's reporting utility. This is a standard operation for the LangWatch CLI to provide feedback and bug reports to the developers.
[COMMAND_EXECUTION]: The skill performs various CLI operations using langwatch to retrieve trace data, monitor configurations, and analytics from the LangWatch platform. These commands are essential for the skill's primary purpose of production debugging.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data retrieved from production traces and documentation, which could potentially contain untrusted content.
Ingestion points: Output from langwatch trace get, langwatch monitor get, and langwatch docs commands, as well as production trace content.
Boundary markers: The skill frequently uses the -o json flag, which provides structured data that can be parsed by the agent rather than treated as direct instructions.
Capability inventory: The ability to execute langwatch CLI commands and perform network queries to the LangWatch API.
Sanitization: The reporting command (langwatch report) includes a local scrubbing step to remove secrets, API keys, and PII from the data before it is sent.

debug-with-langwatch