The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses platform CLI tools such as kubectl, gcloud, and fly to retrieve environment diagnostics and logs. This behavior is expected and necessary for the primary purpose of debugging production infrastructure.
[PROMPT_INJECTION]: The skill ingests untrusted data from application logs and git histories, creating a surface for indirect prompt injection. This risk is inherent to diagnostic tools and is managed by the agent's internal safety protocols and the skill's specific operational context.

vigil-incident