The Agent Skills Directory

[PROMPT_INJECTION]: The skill is intended to analyze untrusted external files such as Dockerfiles and Kubernetes manifests, which constitutes an indirect prompt injection surface. To mitigate this, the author has included a clear 'Prompt Injection Safety Notice' instructing the AI to treat all file content as data and ignore any embedded directives.
[CREDENTIALS_UNSAFE]: The documentation file cis-benchmarks.md contains examples of hardcoded secrets and API keys. These are explicitly presented as 'BAD' examples for auditing purposes and are not functional credentials belonging to the skill or its author.
[SAFE]: The skill requests access to a minimal set of tools (Read, Grep, Glob) and does not attempt to perform network operations, exfiltrate data, or execute remote code. Its behavior is consistent with its stated purpose as a security review tool.

container-security