The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes kubectl to perform cluster diagnostics, including retrieving logs, describing pod status, and executing commands inside containers (e.g., redis-cli, df, stat). It also references local shell scripts (diagnose.sh and collect-logs.sh) intended for automated triage.
[DATA_EXFILTRATION]: While the skill accesses sensitive diagnostic information such as environment variables (OTEL, CORPUS) and database metrics via redis-cli, no network operations were detected that would transmit this data outside the cluster environment.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it ingests untrusted data through kubectl logs and kubectl exec outputs. Malicious content residing in logs or container files could potentially attempt to influence the agent's behavior.
Ingestion points: kubectl logs and kubectl exec outputs (SKILL.md)
Boundary markers: Absent; the skill does not specify delimiters or instructions to ignore embedded commands in the processed data
Capability inventory: Shell command execution via bash and kubectl (SKILL.md)
Sanitization: Absent; output from the cluster is processed without evident filtering or escaping.

debug-buttercup