The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted user-provided error data and stack traces which can contain embedded instructions. This represents an indirect prompt injection surface. The ingestion points are located in SKILL.md under the diagnostic framework section. While the skill possesses capabilities such as Bash and Edit, there are no explicit boundary markers or sanitization logic defined in the skill documentation to mitigate the risk of the agent following instructions hidden within logs.
[COMMAND_EXECUTION]: The documentation includes examples for using system diagnostic tools like systemctl, docker, netstat, lsof, and nc. These tools are recommended for debugging and troubleshooting purposes within the context of the Bash tool. They are standard for system administration and do not constitute malicious behavior in this context.

error-diagnosis