The Agent Skills Directory

[PROMPT_INJECTION]: No instructions were found that attempt to override agent safety guidelines or ignore previous system prompts. The skill is focused on quality evaluation metrics.- [DATA_EXFILTRATION]: The skill does not perform any network operations, HTTP requests, or file system access that could lead to data exposure.- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or secrets were detected. The skill specifically includes a guardrail to identify and reject hardcoded secrets in the content it evaluates.- [REMOTE_CODE_EXECUTION]: There are no references to external scripts, package installations, or remote code downloads.- [COMMAND_EXECUTION]: The skill contains only Markdown instructions and YAML configuration. It does not invoke any shell commands or subprocesses.- [NO_CODE]: The analyzed files consist entirely of Markdown-based prompting instructions and YAML metadata; no executable programming logic (Python, JavaScript, etc.) is included.

output-critic