The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data through the list_logs() tool, which reads service logs that may contain attacker-controlled content. This data is processed alongside high-privilege capabilities such as query_render_postgres and various shell commands, creating a surface for instructions embedded in logs to influence agent actions.
Ingestion points: list_logs() (SKILL.md).
Boundary markers: None specified for log content.
Capability inventory: query_render_postgres (SQL execution), render ssh (terminal access), and multiple render CLI monitoring tools.
Sanitization: No explicit sanitization or instruction-ignoring markers are documented for the data returned by monitoring tools.
[COMMAND_EXECUTION]: The skill utilizes the render CLI for various monitoring tasks, including fetching service status, streaming logs, and establishing SSH connections for debugging. These are standard administrative actions for the service provider.

render-monitor