The Agent Skills Directory

[SAFE]: Analysis across all ten threat categories revealed no malicious patterns or security violations. The skill focuses on standard infrastructure troubleshooting.
[COMMAND_EXECUTION]: The skill provides instructions for using industry-standard tools like docker and kubectl for troubleshooting containerized applications. These include commands for inspecting state, viewing logs, and checking resource usage.
[PROMPT_INJECTION]: Potential indirect prompt injection surfaces were evaluated, specifically the ingestion of container logs and metadata. The skill provides instructions for manual or supervised debugging steps rather than automated processing of untrusted data, which aligns with safe operational practices.
Ingestion points: docker logs, docker cp, docker exec, and kubectl logs as documented in SKILL.md and references/docker-debugging-basics.md.
Boundary markers: None present in the instructions.
Capability inventory: Execution of docker and kubectl binaries on the host system.
Sanitization: Not present; the skill relies on standard output from local container management engines.

container-debugging