claude-code-agent-design
Pass
Audited by Gen Agent Trust Hub on Aug 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of a markdown documentation file (
SKILL.md) that provides expert guidance on agent system design. It contains no executable scripts, automated commands, or external dependencies. - [COMMAND_EXECUTION]: While the guide includes code snippets illustrating shell command execution tools (e.g.,
execute_command), these are provided as static educational examples for developers building their own tools and are not invoked by the skill itself. - [DATA_EXFILTRATION]: The documentation references project context and environment variables in the context of
CLAUDE.mdusage, but this is presented solely as a template for project documentation and does not perform any data access or network transmission. - [PROMPT_INJECTION]: The skill focuses on teaching context engineering and system prompt construction. It does not contain instructions that attempt to override agent safety filters or bypass architectural constraints.
Audit Metadata