The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external observability stacks (Tempo, Loki) and host container logs. Ingestion points: Traces, service logs, and raw container stdout via SSH are identified as input sources in SKILL.md. Boundary markers: The skill lacks explicit delimiters or instructions for the agent to ignore embedded commands within ingested logs. Capability inventory: The agent utilizes task commands to perform network operations and SSH to Dokploy hosts. Sanitization: The skill prescribes redacting PII for public reports but does not specify sanitization of data before it enters the agent context.\n- [COMMAND_EXECUTION]: The skill executes shell-based tasks including SSH connections to Dokploy hosts for log retrieval and Docker operations.\n- [DATA_EXFILTRATION]: The skill facilitates access to sensitive system configurations such as ~/.ssh/config and production log data which may expose runtime metadata or internal identifiers.

debug-trace