The Agent Skills Directory

[SAFE]: The skill does not contain any executable scripts (such as .py, .js, or .sh files). It is comprised entirely of Markdown instructions and YAML configuration files, which do not have the capability to perform system actions directly.
[OBFUSCATION]: The presence of Cyrillic characters in the description (e.g., 'диагностика', 'алертов') represents legitimate Russian language content related to the skill's purpose (diagnostics and alerts). These are not homoglyph attacks as they are entire words in a single script rather than characters substituted into Latin words.
[DATA_EXPOSURE]: Analysis of the YAML and Markdown content shows no hardcoded API keys, tokens, or sensitive system file paths.
[PROMPT_INJECTION]: The instructions focus on defining the SRE role, SLO frameworks, and observability practices. There are no attempts to override safety guardrails or induce the agent to ignore its primary directives.

agency-sre