The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches and processes log data from external VictoriaLogs instances, which could contain malicious instructions designed to manipulate the AI agent.
Ingestion points: Untrusted data enters the agent context through curl responses from various log query and discovery endpoints (e.g., /select/logsql/query).
Boundary markers: The skill instructions do not utilize specific delimiters or warnings to help the agent distinguish between its system instructions and the processed log data.
Capability inventory: The skill utilizes Bash(curl:*) and Bash(jq:*) to perform its tasks, which represent a potential capability surface if the agent were influenced by injected content.
Sanitization: No sanitization, escaping, or filtering of the log content (such as the _msg field) is performed before the data is presented to the agent.

victorialogs-query