The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes phrases such as "ignore previous instructions" and "you are now...". These are documented as examples of patterns to detect during a security audit and are not active instructions intended to override agent behavior. The presence of these strings was correctly flagged by static analysis but they are benign in this context.
[DATA_EXPOSURE]: The framework mentions sensitive paths like ~/.ssh/id_rsa and .env files to illustrate data exfiltration risks that auditors should look for in other skills. This skill does not attempt to access or exfiltrate these files.
[REMOTE_CODE_EXECUTION]: The documentation identifies shell piping (e.g., curl | bash) as a critical risk factor. This is listed in an illustrative table for educational purposes for the auditor.
[NO_CODE]: The skill is comprised entirely of markdown-based instructions, tables, and examples. It does not include any executable scripts, configuration for tool execution, or third-party dependencies.

skill-audit