bootstrap-claude-md
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands like pwd, git, and ls to gather environment context. A potentially destructive rm command is present but is safely managed through mandatory user confirmation.
- [INDIRECT_PROMPT_INJECTION]: By reading local files like package.json and source code to generate summaries, the skill possesses an indirect prompt injection surface. This is considered low risk as the ingestion is required for the skill's primary documentation purpose.
- [SAFE_PRACTICE]: The instructions explicitly prohibit the agent from accessing sensitive directories such as the global Claude configuration folder (~/.claude/), ensuring project analysis remains localized and secure.
Audit Metadata